ES3039109T3 - Methods and materials for assessing homologous recombination deficiency - Google Patents

Abstract

This document provides methods and materials for evaluating samples (e.g., cancer cells) for homology-directed repair (HRD) deficiency or an HRD signature. For example, methods and materials are provided for determining whether a cell (e.g., a cancer cell) contains an HRD signature. Materials and methods are also provided for identifying cells (e.g., cancer cells) with a homology-directed repair (HDR) deficiency, as well as for identifying cancer patients likely to respond to a specific treatment regimen. (Automatic translation using Google Translate, not legally binding)

Description

DESCRIPTION

Methods and materials for assessing homologous recombination deficiency

Background

Cancer is a serious public health problem; in 2009 alone, 562,340 people died from cancer in the United States. (American Cancer Society, Cancer Facts & Figures 2009, available on the American Cancer Society website). One of the major challenges in cancer treatment is discovering relevant and clinically useful characteristics of the patient's cancer and then, based on these characteristics, administering a treatment plan that is best suited to the patient's cancer. Although much progress has been made in this field of personalized medicine, there remains a significant need for better molecular diagnostic tools to characterize patients' cancers. WO2013/130347 describes methods for predicting cancer response to cancer therapies based on the determination and analysis of a chromosomal aberration score. Furthermore, WO2012/027224 discloses the identification of novel methods for defining predictive biomarkers for cancer drug response. Additionally, Popova et al. (Cancer Research 2012, 72:5454) report that ploidy and large-scale genomic instability consistently identify basal-like carcinomas with BRCA1/2 inactivation. Also, an HRD test is known as the Myriad's HRD™ test.

Summary

This document relates to methods and materials involved in the evaluation of samples (e.g., cancer cells or nucleic acids derived from them) with respect to homology-dependent repair (HRD) deficiency (e.g., an HRD signature) based on the detection of particular chromosomal aberrations ("ACs"). For example, this document provides methods and materials for detecting AC regions to determine whether a cell (e.g., a cancer cell) has HRD (e.g., whether it exhibits an HRD signature). This document also provides materials and methods for identifying cancer patients likely to respond to a particular cancer treatment regimen based on the presence, absence, or severity of HRD. Throughout this document, unless otherwise stated, HRD and homology-dependent repair deficiency (HDR) are used interchangeably.

The present invention is defined in the independent claim, to which reference shall now be made. Advantageous embodiments are set forth in the dependent claims. One aspect presents a method for assessing HRD in a cancer cell or DNA (e.g., genomic DNA) derived therefrom. The method may comprise, or essentially consist of, (a) detecting, in a sample or DNA derived therefrom, AC regions (as defined herein) in at least one pair of human chromosomes in the sample or DNA derived therefrom (e.g., any pair of human chromosomes other than a pair of human X/Y sex chromosomes); and (b) determining the number, size (e.g., length), and/or character of such AC regions. The AC regions may be analyzed in several pairs of chromosomes that are representative of the entire genome (e.g., enough chromosomes are analyzed such that the number and size of the AC regions are expected to be representative of the number and size of AC regions in the entire genome).

Several aspects of this disclosure involve the use of a combined analysis of two or more types of AC regions to evaluate (e.g., detect) HRD in a sample. Three types of AC regions useful in such methods include (1) chromosomal regions exhibiting loss of heterozygosity ("LOH regions," as defined herein), (2) chromosomal regions exhibiting telomeric allelic imbalance ("TALI regions," as defined herein), and (3) chromosomal regions exhibiting large-scale transition ("LST regions," as defined herein). AC regions of a particular size, location, or chromosomal character ("indicator AC regions," as defined herein) are particularly useful in the various aspects of the invention described herein.

Therefore, in one respect, a method for evaluating (e.g., detecting) HRD in a sample is described but not claimed herein, comprising (1) determining in the sample the total number of LOH regions of a certain size or character (e.g., "indicator LOH regions," as defined herein); (2) determining in the sample the total number of TAI regions of a certain size or character (e.g., "indicator TAI regions," as defined herein); and (3) evaluating the HRD in the sample based, at least in part, on the determinations made in (1) and (2). (2) determining in the sample the total number of LST regions of a certain size or character (e.g., "indicator LST regions," as defined herein); and (3) evaluating the HRD in the sample based, at least in part, on the determinations made in (1) and (2). A method for evaluating (e.g., detecting) HRD in a sample is further described but not claimed herein, comprising (1) determining the total number of TAI regions of a certain size or character (e.g., "indicator TAI regions," as defined herein) in the sample; (2) determining in the sample the total number of LST regions of a certain size or character (e.g., "indicator LST regions," as defined herein); and (3) evaluating the HRD in the sample based, at least in part, on the determinations made in (1) and (2). The invention provides a method for evaluating (e.g., detecting) HRD in a sample, comprising (1) determining in the sample the total number of LOH regions of a certain size or character (i.e., "indicator LOH regions", as defined herein); (2) determining in the sample the total number of TAI regions of a certain size or character (i.e., "indicator TAI regions", as defined herein); (3) determining in the sample the total number of LST regions of a certain size or character (i.e., "indicator LST regions", as defined herein); and (4) evaluating (e.g., detecting) HRD in the sample based, at least in part, on the determinations made in (1), (2), and (3).

A method for diagnosing the presence or absence of HRD in a patient sample is described but not claimed herein, the method comprising (1) analyzing (e.g., examining) one or more patient samples to determine (e.g., detect) in the sample, the total number of LOH regions of a certain size or character (e.g., indicator "LOH regions," as defined herein); (2) analyzing (e.g., examining) one or more patient samples to determine (e.g., detect) in the sample, the total number of TAI regions of a certain size or character (e.g., indicator "TAI regions," as defined herein); and (3)(a) diagnosing the presence of HRD in a patient sample where the number in (1) and/or the number in (2) exceeds some reference; or (3)(b) diagnosing the absence of HRD in a patient sample where neither the number of (1) nor the number of (2) exceeds any reference. A method for diagnosing the presence or absence of HRD in a patient sample is also described but not claimed herein, the method comprising (1) analyzing (e.g., examining) one or more patient samples to determine (e.g., detect) in the sample the total number of LOH regions of a certain size or character (e.g., indicator "LOH regions," as defined herein); (2) analyzing (e.g., examining) one or more patient samples to determine (e.g., detect) in the sample the total number of LST regions of a certain size or character (e.g., indicator "LST regions," as defined herein); and (3)(a) diagnosing the presence of HRD in a patient sample where the number of (1) and/or the number of (2) exceeds any reference; or (3)(b) diagnosing the absence of HRD in a patient sample where neither the number of (1) nor the number of (2) exceeds any reference. A method for diagnosing the presence or absence of HRD in a patient sample is further described but not claimed herein, the method comprising (1) analyzing (e.g., examining) one or more patient samples to determine (e.g., detect) in the sample, the total number of TAI regions of a certain size or character (e.g., indicator "TAI regions," as defined herein); (2) analyzing (e.g., examining) one or more patient samples to determine (e.g., detect) in the sample, the total number of LST regions of a certain size or character (e.g., indicator "LST regions," as defined herein); and (3)(a) diagnosing the presence of HRD in a patient sample where the number of (1) and/or the number of (2) exceeds any reference; or (3)(b) diagnosing the absence of HRD in a patient sample where neither the number of (1) nor the number of (2) exceeds any reference. In another respect, the disclosure provides a method for diagnosing the presence or absence of HRD in a patient sample, the method comprising (1) analyzing (e.g., examining) one or more patient samples to determine (e.g., detect) in the sample, the total number of LOH regions of a certain size or character (e.g., indicator "LOH regions," as defined herein); (2) analyze (e.g., examine) one or more patient samples to determine (e.g., detect) in the sample, the total number of TAI regions of a certain size or character (e.g., indicator "TAI regions", as defined herein); (3) analyze (e.g., examine) one or more patient samples to determine (e.g., detect) in the sample, the total number of LST regions of a certain size or character (e.g., indicator "LST regions", as defined herein); and (3)(a) diagnose the presence of HRD in a patient sample where the number of (1), the number of (2) and/or the number of (3) exceeds any reference; or (3)(b) diagnose the absence of HRD in a patient sample where none of the numbers of (1), (2) or (3) exceeds any reference.

According to the subclaims, the claimed method may involve the use of an average (e.g., arithmetic mean) of three types of AC regions to assess (e.g., detect) HRD in a sample. Three types of AC regions useful in such methods include (1) chromosomal regions exhibiting loss of heterozygosity ("LOH regions," as defined herein), (2) chromosomal regions exhibiting telomeric allelic imbalance ("TALI regions," as defined herein), and (3) chromosomal regions exhibiting large-scale transition ("lSt regions," as defined herein). AC regions of a certain size or character (e.g., "indicator AC regions," as defined herein) may be particularly useful in the various aspects of the method described herein. Therefore, in some embodiments, the invention provides a method for evaluating (e.g., detecting) HRD in a sample, comprising (1) determining in the sample the total number of LOH regions of a certain size or character (i.e., "indicator LOH regions," as defined herein); (2) determining in the sample the total number of TAI regions of a certain size or character (i.e., "indicator TAI regions," as defined herein); (3) determining in the sample the total number of LST regions of a certain size or character (i.e., "indicator LST regions," as defined herein); (4) calculating the average (e.g., arithmetic mean) of the determinations made in (1), (2), and (3); and (5) evaluating the HRD in the sample based, at least in part, on the calculated average (e.g., arithmetic mean) made in (4).

The assessment (e.g., detection) of HRD is based on a score derived from (e.g., representing or corresponding to) the detected AC regions ("AC region score," as defined herein). Scores are described in greater detail herein. HRD is detected if an AC region score from a sample exceeds a certain threshold (e.g., a baseline or initial AC region score), and optionally, HRD is not detected if the sample's AC region score does not exceed a certain threshold (e.g., a baseline or initial AC region score, which in some implementations may be the same as the positive detection threshold). Those skilled in the art will readily appreciate that, in this disclosure, scores can be designed in the opposite direction (e.g., HRD is detected if the AC region score is below a certain threshold and not detected if the score is above a certain threshold).

As disclosed herein, the AC region score is a combination of scores derived from (e.g., representing or corresponding to) two or more of (1) the detected LOH regions ("LOH region score", as defined herein), (2) the detected TAI regions ("TAI region score", as defined herein), and/or (3) the detected LST regions ("LST region score", as defined herein).

The LOH region score, the TAI region score, and the LST region score are combined as follows to give an AC region score:

AC region score = A*(LOH region score) B*(TAI region score) C*(LST region score)

For example, the LOH region score, the TAI region score, and the LST region score are combined as follows to give an AC region score:

AC region score = 0.21*(LOH region score) 0.67*(TAI region score)

+ 0.12*(LST region score)

As disclosed herein, the AC region score may be a combination of scores derived from (e.g., representing or corresponding to) the average (e.g., arithmetic mean) of (1) the detected LOH regions ("LOH region score", as defined herein), (2) the detected TAI regions ("TAI region score", as defined herein), and/or (3) the detected LST regions ("LST region score", as defined herein) to give an AC region score:

A* (LOH region score) B* (TAI region score) C* (AC region score =_____________________________ of the LS T region)_____________________________

3

A method for predicting the status of the BRCA1 and BRCA2 genes in a sample is described but not claimed herein. This method is analogous to methods described above and differs in that the determination of AC regions, LOH regions, TAI regions, LST regions, or scores incorporating them, is used to assess (e.g., detect) a BRCA1 and/or BRCA2 deficiency in the sample. In some claims, the claimed method may further comprise predicting the response of a cancer patient to a cancer treatment regimen comprising a DNA-damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and/or a PARP (poly ADP ribose polymerase) inhibitor. This method is analogous to the methods described above and differs in that the determination of AC regions, LOH regions, TAI regions, LST regions, or scores incorporating them, is used to predict the likelihood that a cancer patient will respond to the cancer treatment regimen. The patients may be treatment-naïve. A cancer treatment method is further described but not claimed herein. This method is analogous to the methods described above and differs in that a particular treatment regimen (recommended, prescribed, etc.) is administered based, at least in part, on the determination of AC regions, LOH regions, TAI regions, LST regions, or scores incorporating them. For the avoidance of doubt, the treatment methods are not included in the claimed invention. The use of one or more drugs selected from the group consisting of DNA-damaging agents, anthracyclines, topoisomerase I inhibitors, and PAR inhibitors in the manufacture of a drug useful for the treatment of cancer in a patient who has been identified as having (or having had) a cancer cell determined to have HRD (e.g., an HRD signature) as described herein is further described but not claimed herein. A method for assessing the presence of a mutation in a gene of an HDR pathway in a sample is further described but not claimed herein. This method is analogous to the methods described above and differs in that the determination of AC regions, LOH regions, TAI regions, LST regions, or scores incorporating them, is used to detect (or not detect) the presence of a mutation in a gene of an HDR pathway.

A method for evaluating a patient is further described but not claimed herein. The method comprises, or essentially consists of, (a) determining whether the patient has (or has had) cancer cells with more than one reference number of AC regions (or, e.g., an AC region score exceeding a reference AC region score); and (b)(1) diagnosing the patient as having HRD cancer cells if the patient is determined to have (or has had) cancer cells with more than one reference number of AC regions (or, e.g., an AC region score exceeding a reference AC region score); or (b)(2) diagnosing the patient as not having HRD cancer cells if the patient is determined not to have (or has not had) cancer cells with more than one reference number of AC regions (or, e.g., the patient does not have (or has not had) cancer cells with an AC region score exceeding a reference AC region score).

The use of a plurality of oligonucleotides capable of hybridizing with a plurality of polymorphic regions of human genomic DNA is further described but not claimed herein, in the manufacture of a diagnostic kit useful for determining the total number or combined length of AC regions in at least one pair of chromosomes (or DNA derived therefrom) in a sample obtained from a cancer patient, and for detecting (a) HRD or the likelihood of HRD (e.g., a DRH signature) in the sample, (b) deficiency (or the likelihood of deficiency) in the sample of a BRCA1 or BRCA2 gene, or (c) an increased likelihood of the cancer patient responding to a cancer treatment regimen comprising a DNA-damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, or a PARP inhibitor.

A system for detecting HRD (e.g., an HRD signature) in a sample is also described but not claimed herein. The system comprises, or essentially consists of, (a) a sample analyzer configured to produce a plurality of signals on the genomic DNA of at least one pair of human chromosomes (or DNA derived therefrom) in the sample, and (b) a computer subsystem programmed to calculate, based on the plurality of signals, the number or combined length of AC regions on at least one pair of human chromosomes. The computer subsystem can be programmed to compare the number or combined length of AC regions with a reference number to detect (a) the probability of HRD (e.g., a DRH signature) in the sample, (b) deficiency (or probability of deficiency) in the sample of a BRCA1 or BRCA2 gene, or (c) an increased probability that the cancer patient will respond to a cancer treatment regimen comprising a DNA-damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, or a PARP inhibitor. The system can include an output module configured to display (a), (b), or (c). The system can also include an output module configured to display a recommendation regarding the use of the cancer treatment regimen.

A computer program product embodied in a computer-readable medium is also described but not claimed herein. This product, when executed on a computer, provides instructions for detecting the presence or absence of any AC region along one or more human chromosomes other than the human sex chromosomes X and Y (the AC regions optionally being indicator AC regions); and for determining the total number or combined length of AC regions on the one or more pairs of chromosomes. The computer program product may include other instructions.

A diagnostic kit is further described but not claimed herein. The kit comprises, or essentially consists of, at least 500 oligonucleotides capable of hybridizing with a plurality of polymorphic regions of human genomic DNA (or DNA derived therefrom); and a software product disclosed herein. The software product may be incorporated on a computer-readable medium which, when executed on a computer, provides instructions for detecting the presence or absence of any AC region along one or more human chromosomes other than the human sex chromosomes X and Y (the AC regions optionally being indicator AC regions); and determining the total number or combined length of the AC regions on the one or more pairs of chromosomes. The software product may include other instructions.

In some embodiments of the method of the invention, any one or more of the following may be applied as appropriate. The AC regions may be determined on at least two, five, ten, or 21 pairs of human chromosomes. The cancer cell may be an ovarian, breast, lung, or esophageal cancer cell. The reference may be 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, or 20 or higher. The at least one pair of human chromosomes may exclude human chromosome 17. The DNA-damaging agent may be cisplatin, carboplatin, oxaliplatin, or picoplatin; the anthracycline may be epirubinic or doxorubicin; the topoisomerase 1 inhibitor may be campothecin, topotecan, or irinotecan; or the PARP inhibitor may be iniparib, olaparib, or velaparib. The patient may be a patient without prior treatment.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as understood by a person skilled in the art to which this disclosure pertains. Although methods and materials similar to or equivalent to those described herein may be used to implement the invention, suitable methods and materials are described below.

Details of one or more embodiments of the invention are set forth below in the accompanying drawings and description. Other features, objects, and advantages of the invention will become apparent from the description, the drawings, and the claims.

Description of the drawings

Figure 1 shows graphs representing allelic doses of breast cancer cells from a fresh frozen sample of a breast cancer patient along a chromosome, determined using an SNP array (top) and high-throughput sequencing (bottom).

Figure 2 shows graphs representing allelic doses of breast cancer cells from a formalin-fixed, paraffin-embedded (FFPE) sample from a breast cancer patient along a chromosome, determined using an SNP array (top) and high-throughput sequencing (bottom). Figure 3 is a flowchart of an example process for evaluating a cell's genome (e.g., a cancer cell) against an HRD signature.

Figure 4 is a diagram of an example of a computing device and a portable computing device that can be used to implement the techniques described in this document.

Figure 5A shows LOH region scores in IHC (immunohistochemical) subtypes of breast cancer. The top three panels are BRCA1/2-deficient samples. The bottom panel is BRCA1/2-unaltered samples.

Figure 5B shows TAI region scores in IHC subtypes of breast cancer. The top three panels are BRCA1/2-deficient samples. The bottom panel is BRCA1/2-unaltered samples. Figure 6 shows the correlation between LOH and TAI region scores. Correlation coefficient = 0.69. X-axis: LOH score; Y-axis: TAI score; red dots: unaltered samples; blue dots (with an "X" superimposed): BRCA1/2-deficient samples. The area under the dots is proportional to the number of samples with that combination of LOH and TAI scores. p = 10⁻³⁹.

Figure 7A shows LOH region scores from the patients analyzed in Example 2 of this document. The top three panels are BRCA1/2 deficient samples. The bottom panel shows unaffected BRCA1/2 samples.

Figure 7B shows TAI region scores from the patients analyzed in Example 2 of this document. The top three panels are BRCA1/2 deficient samples. The bottom panel shows unaffected BRCA1/2 samples.

Figure 7C shows LST region scores from the patients analyzed in Example 2 of this document. The top three panels are BRCA1/2 deficient samples. The bottom panel shows unaffected BRCA1/2 samples.

Figure 7D shows LOH versus TAI scores for patients analyzed in Example 2 of this document. X-axis: LOH score; Y-axis: TAI score; red dots: unaffected samples; blue dots (with an "X" superimposed): BRCA1/2 deficient samples. The area under the dots is proportional to the number of samples with that combination of LOH and TAI scores.

Figure 7 shows LOH versus LST scores from patients analyzed in Example 2 of this document. X-axis: LOH score; Y-axis: LST score; red dots: unaffected samples; blue dots (with an "X" superimposed): BRCA1/2 deficient samples. The area under the dots is proportional to the number of samples with that combination of LOH and LST scores.

Figure 7F shows TAI versus LST scores from patients analyzed in Example 2 of this document. X-axis: TAI score; Y-axis: LST score; red dots: unaffected samples; blue dots (with an "X" superimposed): BRCA1/2 deficient samples. The area under the dots is proportional to the number of samples with that combination of TAI and LST scores.

Figure 8 is a graph representing the number of LOH regions longer than 15 Mb and shorter than the entire chromosome in ovarian cancer cell samples with somatic BRCA mutations, germline BRCA mutations, low BRCA1 expression, or unaltered BRCA (normal BRCA). The size of the circles is proportional to the number of samples with that number of LOH regions.

Figure 9 illustrates HRD-LOH scores in BRCA 1/2 deficient (mutated or methylated) (top panel) and unaltered (bottom panel) samples in a full breast cancer cohort.

Figure 9 illustrates HRD-TAI scores in BRCA1/2 deficient (mutated or methylated) (top panel) and unaltered (bottom panel) samples in a full breast cancer cohort.

Figure 9 illustrates HRD-LST scores in BRCA1/2 deficient (mutated or methylated) (top panel) and unaltered (bottom panel) samples in a full breast cancer cohort.

Figure 10 illustrates a combined average (e.g., arithmetic mean) HRD score (Y-axis) stratified by the Miller-Payne score (horizontal axis) in combined cisplatin-1 and cisplatin-2 cohorts.

Figure 11 illustrates a Spearman correlation of 3 different measures of HR deficiency (homologous recombination). The panels above the diagonal show correlation. The panels diagonally show density plots.

Figure 12 illustrates associations of clinical variables with the combined HRD score.

Figure 13 illustrates associations of clinical variables with BRCA1/2 deficiency. The upper panels and the lower left panel show the proportion of patients with BRCA1/2 deficiency in each category of breast cancer grade, stage, and type. The width of each bar is proportional to the number of patients in each category. The lower right panel shows an estimate of the conditional density of BRCA1/2 deficiency by age.

Figure 14 illustrates a high HRD determination with a reference score > 42.

Figure 15 illustrates a histogram showing the distribution of HRD scores in a cisplatin cohort. The four columns on the left represent low HRD, and the five columns on the right, with reference scores > 42, represent low HRD.

Figure 16 illustrates the distribution of HRD scores across the pCR, RCB-I, RCB-II, and RCB-III response classes. The boxes represent the interquartile range (IQR) of the scores, with a horizontal line at the median. The dotted line at the value 42 represents the HRD threshold between low and high scores.

Figure 17 illustrates a response curve for the quantitative HRD score. The curve is modeled using generalized logistic regression. The shaded cells indicate the probability of response in samples with homologous recombination deficiency (HR) versus samples without such deficiency.

Figure 18 illustrates HRD scores for the individual components of HRD (LOH, TAI, and LST).

Detailed description

A method for assessing HRD in a cancer cell or DNA (e.g., genomic DNA) derived from it is disclosed herein. The method may comprise, or essentially consist of, (a) detecting, in a sample or DNA derived from it, regions of AC in at least one pair of human chromosomes or DNA derived from them; and (b) determining the number, size (e.g., length), and/or character of such AC regions.

As used herein, "chromosomal aberration" or "AC" means a somatic change in the chromosomal DNA of a cell that falls into at least one of three overlapping categories: LOH, TAI, or LST. Polymorphic loci within the human genome (e.g., single nucleotide polymorphisms [SNPs]) are generally heterozygous within an individual's germline, as that individual typically receives one copy from the biological father and one copy from the biological mother. However, from a somatic perspective, this heterozygosity can change (through mutation) to homozygosity. This change from heterozygosity to homozygosity is called loss of heterozygosity (LOH). LOH can result from several mechanisms. For example, in some cases, a locus on a chromosome may be deleted in a somatic cell. The locus that remains present on the other chromosome (the other non-sex male chromosome) is an LOH locus because there is only one copy (instead of two copies) of that locus present in the genome of the affected cells. This type of LOH event results in a copy number reduction. In other cases, a locus on a A chromosome (e.g., a male non-sex chromosome) in a somatic cell can be replaced by a copy of that locus from the other chromosome, thereby eliminating any heterozygosity that may have been present within the replaced locus. In such cases, the locus that remains present on each chromosome is an LOH locus and may be referred to as a neutral copy LOH locus. LOH and its use in HRD determination are described in detail in International Application No. PCT/US2011/040953 (published as WO/2011/160063).

A broader class of chromosomal aberration, which includes LOH, is allele imbalance. Allele imbalance occurs when the relative number of copies (i.e., copy ratio) at a particular locus in somatic cells differs from that in the germline. For example, if the germline has one copy of allele A and one copy of allele B at a particular locus, and a somatic cell has two copies of A and one copy of B, there is allele imbalance at the locus because the copy ratio in the somatic cell (2:1) differs from that in the germline (1:1). LOH is an example of allele imbalance since the somatic cell has a copy ratio (1:0 or 2:0) that differs from that in the germline (1:1). But allele imbalance encompasses more types of chromosomal aberration, e.g., germline 2:1 to somatic 1:1; germline 1:0 to somatic 1:1; germline 1:1 to somatic 2:1, etc. The analysis of regions of allelic disequilibrium spanning chromosome telomeres is particularly useful in the invention. Therefore, a "telomeric allelic disequilibrium region" or "TAI Region" is defined as a region with allelic disequilibrium that (a) extends to one of the subtelomeres and (b) does not cross the centromere. TAI and its use in determining h Rd are described in detail in U.S. patent applications serial numbers 13/818,425 (published as US20130281312A1) and 14/466,208 (published as US20150038340A1).

A class of chromosomal aberrations that is even broader, encompassing LOH and TAI, is referred to herein as large-scale transition ("LST"). LST refers to any somatic copy number transition (i.e., breakpoint) along a chromosome when it lies between two regions of at least a certain minimum length (e.g., at least 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 or 20 or more megabases) after filtering out regions shorter than a certain maximum length (e.g., 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4 or more megabases). For example, if after filtering out regions shorter than 3 megabases, the somatic cell has a 1:1 copy number for, say, at least 10 megabases, and then a breakpoint transition occurs in a region of, say, at least 10 megabases with a 2:2 copy number, this is an LST. An alternative way to define the same phenomenon is as an LST region, which is a genomic region with a copy number of

minimum length (e.g., at least 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 megabases) delimited by breakpoints (i.e., transitions) where the copy number changes to another region also at least of this minimum length. For example, if after filtering out regions shorter than 3 megabases, the somatic cell has a region of at least 10 megabases with a 1:1 copy number delimited on one side by a breakpoint transition to a region of, e.g., at least 10 megabases with a 2:2 copy number, and delimited on the other side by a breakpoint transition to a region of, e.g., For example, at least 10 megabases with a copy number of 1:2, then this is two LSTs. It should be noted that this is broader than allelic imbalance because such a change in copy number would not be considered allelic imbalance (since the 1:1 and 2:2 copy ratios are the same, i.e., there has been no change in the copy ratio). The LST and its use in determining HRD are described in detail in U.S. patent application serial no. 14/402,254 (published as US20150140122A1).

Different LST score cutoffs can be used for "near-diploid" and "near-tetraploid" tumors to separate unaltered and BRCA1/2-deficient samples. The LST score sometimes increases with ploidy in both unaltered and deficient samples. As an alternative to using ploidy-specific cutoffs, some implementations may employ a modified LST score adjusted for ploidy: LSTm = LST - kP, where P is ploidy and k is a constant. Based on a multivariate logistic regression analysis with deficiency as an outcome and LST and P as independent variables, k = 15.5 provided the best separation between unaltered and deficient samples (although a person skilled in the art might consider other values for k).

Chromosomal aberrations can extend across numerous loci to define a chromosomal aberration region, hereinafter referred to as an "AC region." Such AC regions can be of any length (e.g., from less than approximately 1.5 Mb to the entire length of the chromosome). An abundance of large AC regions ("indicator AC regions") indicates a deficiency in a cell's homology-dependent repair (HDR) mechanism. The definition of an AC region, and therefore what constitutes an "indicator" region, for each type of AC (e.g., LOH, TAI, LST) depends on the particular nature of the AC. For example, an "LOH region" means at least some minimum number of consecutive loci exhibiting an LOH or some minimum stretch of genomic DNA containing consecutive loci exhibiting an LOH. Furthermore, a "TAI region" means at least some minimum number of consecutive loci exhibiting allelic imbalance that extend from the telomere to the rest of the chromosome (or some minimum segment of genomic DNA extending from the telomere to the rest of the chromosome that has consecutive loci exhibiting allelic imbalance). The LST is already defined as a region of genomic DNA of at least some minimum size; therefore, in this document, "LST" and "LST region" are used interchangeably to refer to a minimum number of consecutive loci (or some minimum segment of genomic DNA) that have the same copy number, delimited by a breakpoint or transition from that copy number to a different one.

In some embodiments, an AC region (whether an LOH region, a TAI region, or an LST region) is an indicator AC region (whether an indicator LOH region, an indicator TAI region, or an indicator LST region) if it has a length of at least 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 25, 30, 35, 40, 45, 50, 60, 70, 80, 90, or 100 megabases or greater. In some embodiments, the indicator LOH regions are LOH regions that are longer than approximately 1.5, 5, 12, 13, 14, 15, 16, 17 or more (preferably 14, 15, 16 or more, most preferably 15 or more) megabases but shorter than the entire length of the respective chromosome within which the LOH region is located. Alternatively or additionally, the total combined length of such indicator LOH regions may be determined. In some embodiments, indicator TAI regions are allelic imbalance TAI regions that (a) extend into one of the subtelomeres, (b) do not cross the centromere, and (c) are longer than 1.5, 5, 12, 13, 14, 15, 16, 17, or more (preferably 10, 11, 12, or more, most preferably 11 or more) megabases. Alternatively or additionally, the combined total length of such indicator TAI regions may be determined. Since the LST concept already implies regions of some minimum size (this minimum size being determined based on their ability to differentiate HRD from unaltered HDR samples), indicator LST regions, such as those used herein, are the same as LST regions. Furthermore, an LST region score may be derived from the number of regions exhibiting LST as described above or from the number of LST breakpoints. In some realizations, the minimum length of the region of the number of

13, 14, 15, 16, 17, 18, 19 or 20 megabases (preferably 8, 9, 10, 11 or more megabases, more preferably 10 megabases) and the maximum region that remains unfiltered is less than 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4 or less megabases (preferably 2, 2.5, 3, 3.5 or 4 or less megabases, more preferably less than 3 megabases).

As used herein, a sample has an "HRD signature" if such sample has a number of indicator AC regions (as described herein) or an AC region score (as described herein) that exceeds a reference as described herein, wherein a number or score that exceeds such reference indicates a homologous recombination deficiency.

Therefore, the invention generally involves the detection and quantification of indicator AC regions in a sample to determine whether the cells in the sample (or the cells from which the sample DNA is derived) have a HRD signature. Often, this comprises comparing the number of indicator AC regions (or a test value or score derived from or calculated therefrom and corresponding to that number) with a baseline or reference number (or score).

The various aspects of this disclosure involve the use of a combined analysis of two or more types of AC regions (including two or more types of indicator AC regions) to evaluate (e.g., detect, diagnose) HRD in a sample. In one aspect, the invention provides a method for evaluating (e.g., detecting, diagnosing) HRD in a sample, comprising (1) determining in the sample the total number (or combined length) of indicator LOH regions; (2) determining in the sample the total number of indicator TAI regions; (3) determining in the sample the total number (or combined length) of indicator LST regions; and (4) determining in the sample the presence or absence (e.g., detect, diagnose) of HRD based, at least in part, on the determinations made in (1), (2), and (3).

The various aspects of the present invention comprise the use of a combined analysis of the averages of three different AC regions to evaluate (e.g., detect, diagnose) HRD in a sample. Therefore, in one aspect, the invention provides a method for evaluating (e.g., detecting, diagnosing) HRD in a sample, which may comprise (1) determining in the sample the total number of LOH regions of a certain size or character (e.g., "indicator LOH regions," as defined herein); (2) determining in the sample the total number of TAI regions of a certain size or character (i.e., "indicator TAI regions," as defined herein); (3) determining in the sample the total number of LST regions of a certain size or character (i.e., "indicator LST regions," as defined herein); (4) calculating the average (e.g., arithmetic mean) of the determinations made in (1), (2), and (3); and (5) evaluating the HRD in the sample based, at least in part, on the calculated average (e.g., arithmetic mean) made in (4).

As used herein, "AC region score" means a test value or score derived from (e.g., representing or corresponding to) indicator AC regions detected in a sample (e.g., a test score or value derived from or calculated from the number of indicator AC regions detected in a sample). Similarly, as used herein, "LOH region score" is a subset of AC region scores and means a test value or score derived from (e.g., representing or corresponding to) indicator LOH regions detected in a sample (e.g., a test score or value derived from or calculated from the number of indicator LOH regions detected in a sample), and so on for TAI region score and LST region score. In some embodiments, such a score may simply be the number of indicator AC regions detected in a sample. In some implementations, the scoring is more complicated, taking into account the lengths of each indicator AC region or a subset of detected indicator AC regions.

A method for assessing (e.g., detecting, diagnosing) HRD in a sample is disclosed but not claimed herein, comprising (1) determining a LOH region score for the sample; (2) determining a TAI region score for the sample; (3) determining an LST region score for the sample; and (4)(a) detecting (or diagnosing) HRD in the sample based, at least in part, on the LOH region score exceeding a reference, the TAI region score exceeding a reference, or the LST region score exceeding a reference; or optionally (4)(b) detecting (or diagnosing) an absence of HRD in the sample based, at least in part, on the LOH region score not exceeding a reference, the TAI region score not exceeding a reference, and the LST region score not exceeding a reference.

As disclosed herein, the AC region score is a combination of scores derived from or calculated from (e.g., representing or corresponding to) (1) the detected LOH regions ("LOH region score", as defined herein), (2) the detected TAI regions ("TAI region score", as defined herein), and/or (3) the detected LST regions ("LST region score", as defined herein).

As disclosed in this document, the LOH region score, the TAI region score, and the LST region score are combined as follows to give an AC region score:

CA region score - A*(LOH region score) B*(TAI region score)

C*(LST region score)

For example, the LOH region score, the TAI region score, and the LST region score are combined as follows to give an AC region score:

AC region score - 0.21*(LOH region score) 0.67*(TAI region score)

0.12*(LST region score)

EITHER

AC region score = [0.24]*(LOH region score) [0.65]*(TAI region score) [0.11]*(LST region score)

EITHER

AC region score = [0.11]*(LOH region score) [0.25]*(TAI region score) [0.12]*(LST region score)

Optionally, the AC region score is a combination of scores derived from (e.g., representing or corresponding to) the average (e.g., arithmetic mean) of (1) the detected LOH regions ("LOH region score", as defined herein), (2) the detected TAI regions ("TAI region score", as defined herein), and (3) the detected LST regions ("LST region score", as defined herein) to give a calculated AC region score as follows:

A*(LOH region score) B*(TAI region score) AC =_____________________C*(LST region score)_____________________

3

One or more of these coefficients (i.e., A, B, or C, or any combination thereof) is 1, and in some realizations all three coefficients (i.e., A, B, and C) are 1. Therefore, in some realizations, the AC region score = (LOH region score) (TAI region score) (LST region score), where the LOH region score is the number of indicator LOH regions (or the total LOH length), the TAI region score is the number of indicator TAI regions (or the total TAI length), and the LST region score is the number of indicator LST regions (or the total LST length).

As described in this document, A is between 0.9 and 1, 0.9 and 0.99, 0.9 and 0.95, 0.85 and 0.95, 0.86 and 0.94, 0.87 and 0.93, 0.88 and 0.92, 0.89 and 0.91, 0.85 and 0.9, 0.8 and 0.95, 0.8 and 0.9, 0.8 and 0.85, 0.75 and 0.99, 0.75 and 0.95, 0.75 and 0.9, 0.75 and 0.85, or between 0.75 and 0.8. As described herein, B is between 0.40 and 1, 0.45 and 0.99, 0.45 and 0.95, 0.55 and 0.8, 0.55 and 0.7, 0.55 and 0.65, 0.59 and 0.63, or between 0.6 and 0.62. As described herein, C is, where applicable, between 0.9 and 1, 0.9 and 0.99, 0.9 and 0.95, 0.85 and 0.95, 0.86 and 0.94, 0.87 and 0.93.

0.88 and 0.92, 0.89 and 0.91, 0.85 and 0.9, 0.8 and 0.95, 0.8 and 0.9, 0.8 and 0.85, 0.75 and 0.99, 0.75 and 0.95, 0.75 and 0.9, 0.75 and

0.85, or between 0.75 and 0.8. As described herein, D is, where applicable, between 0.9 and 1.

0.9 and 0.99, 0.9 and 0.95, 0.85 and 0.95, 0.86 and 0.94, 0.87 and 0.93, 0.88 and 0.92, 0.89 and 0.91, 0.85 and 0.9, 0.8 and 0.95, 0.8 and 0.9, 0.8 and 0.85, 0.75 and 0.99, 0.75 and 0.95, 0.75 and 0.9, 0.75 and 0.85, or between 0.75 and 0.8.

In some aspects of disclosure, A is between 0.1 and 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5,

4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 0.2 and 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4,

4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 0.3 and 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5,

6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 0.4 and 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9,

10, 11, 12, 13, 14, 15 or 20; or between 0.5 and 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13,

14, 15, or 20; or between 0.6 and 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between

0.7 and 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 0.8 and 0.9, 1, 1.5, 2, 2.5,

3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 0.9 and 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11,

12, 13, 14, 15 or 20; or between 1 and 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 1.5 and 2,

2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 2 and 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13,

14, 15 or 20; or between 2.5 and 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 3 and 3.5, 4, 4.5, 5, 6, 7, 8, 9,

10, 11, 12, 13, 14, 15 or 20; or between 3.5 and 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 4 and 4.5, 5, 6, 7, 8,

9, 10, 11, 12, 13, 14, 15 or 20; or between 4, 5 and 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 5 and 6, 7, 8, 9, 10, 11,

12, 13, 14, 15 or 20; or between 6 and 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 7 and 8, 9, 10, 11, 12, 13, 14, 15 or 20; or

between 8 and 9, 10, 11, 12, 13, 14, 15 or 20; or between 9 and 10, 11, 12, 13, 14, 15 or 20; or between 10 and 11, 12, 13, 14, 15 or 20;

or between 11 and 12, 13, 14, 15 or 20; or between 12 and 13, 14, 15 or 20; or between 13 and 14, 15 or 20; or between 14 and 15, or 20; or between 15 and 20; B is between 0.1 and 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 0.2 and 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13,

14, 15 or 20; or between 0.3 and 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15

or 20; or between 0.4 and 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between

0.5 and 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20; or between 0.6 and 0.7, 0.8,

0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 0.7 and 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5,

4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 0.8 and 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11,

12, 13, 14, 15 or 20; or between 0.9 and 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 1 and

1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 1.5 and 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10,

11, 12, 13, 14, 15 or 20; or between 2 and 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 2.5 and 3, 3.5,

4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 3 and 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; either

between 3.5 and 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 4 and 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20;

or between 4.5 and 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 5 and 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 6 and

7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 7 and 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 8 and 9, 10, 11, 12, 13, 14, 15

or 20; or between 9 and 10, 11, 12, 13, 14, 15 or 20; or between 10 and 11, 12, 13, 14, 15 or 20; or between 11 and 12, 13, 14, 15 or 20;

or between 12 and 13, 14, 15 or 20; or between 13 and 14, 15 or 20; or between 14 and 15, or 20; or between 15 and 20; C is, where applicable, between 0.1 and 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,

15 or 20; or between 0.2 and 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15

or 20; or between 0.3 and 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; either

between 0.4 and 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 0.5 and

0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20; or between 0.6 and 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20; or between 0.7 and 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5,

5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 0.8 and 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13,

14, 15 or 20; or between 0.9 and 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 1 and 1.5, 2,

2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 1.5 and 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11,

12, 13, 14, 15 or 20; or between 2 and 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 2.5 and 3, 3.5, 4,

4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 3 and 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between

3.5 and 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 4 and 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between

4.5 and 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 5 and 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 6 and 7, 8, 9,

10, 11, 12, 13, 14, 15 or 20; or between 7 and 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 8 and 9, 10, 11, 12, 13, 14, 15 or 20; or

between 9 and 10, 11, 12, 13, 14, 15 or 20; or between 10 and 11, 12, 13, 14, 15 or 20; or between 11 and 12, 13, 14, 15 or 20; or between

12 and 13, 14, 15 or 20; or between 13 and 14, 15 or 20; or between 14 and 15, or 20; or between 15 and 20; and D is, where applicable, between 0.1 and 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20;

or between 0.2 and 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or

between 0.3 and 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between

0.4 and 0.5, 0.6, 0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 0.5 and 0.6,

0.7, 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 20; or between 0.6 and 0.7, 0.8, 0.9, 1, 1.5,

2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 0.7 and 0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5,

6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 0.8 and 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,

15 or 20; or between 0.9 and 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 1 and 1.5, 2, 2.5, 3,

3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 1.5 and 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13,

14, 15 or 20; or between 2 and 2.5, 3, 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 2.5 and 3, 3.5, 4, 4.5, 5, 6,

7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 3 and 3.5, 4, 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 3.5 and 4,

4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 4 and 4.5, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 4.5 and

5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 5 and 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 6 and 7, 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 7 and 8, 9, 10, 11, 12, 13, 14, 15 or 20; or between 8 and 9, 10, 11, 12, 13, 14, 15 or 20; or between 9 and 10, 11, 12, 13, 14, 15 or 20; or between 10 and 11, 12, 13, 14, 15 or 20; or between 11 and 12, 13, 14, 15 or 20; or between 12 and 13, 14, 15 or 20; or between 13 and 14, 15 or 20; or between 14 and 15, or 20; or between 15 and 20. In some embodiments, A, B and/or C are within the rounding of any of these values (e.g., A is between 0.45 and 0.54, etc.).

Therefore, a method for assessing (e.g., detecting, diagnosing) HRD in a sample is disclosed herein, comprising (1) determining an LOH region score for the sample; (2) determining a TAI region score for the sample; (3) determining an LST region score for the sample; and (4)(a) detecting (or diagnosing) HRD in the sample based, at least in part, on a combination of the LOH region score, the TAI region score, and the LST region score (e.g., a combined AC region score) that exceeds a reference; or optionally (4)(b) detecting (or diagnosing) an absence of HRD in the sample based at least in part on the LOH region score, the TAI region score, and the LST region score (e.g., a combined AC region score) that does not exceed a reference.

Therefore, in some embodiments the invention provides a method for evaluating (e.g., detecting, diagnosing) HRD in a sample, comprising (1) determining in the sample the total number of LOH regions of a certain size or character (i.e., "indicator LOH regions", as defined herein); (2) determining in the sample the total number of TAI regions of a certain size or character (i.e., "indicator TAI regions", as defined herein); (3) determining in the sample the total number of LST regions of a certain size or character (i.e., "indicator LST regions", as defined herein); (4) calculating the average (e.g., arithmetic mean) of the determinations made in (1), (2), and (3); and (5) evaluating the HRD in the sample based, at least in part, on the calculated average (e.g., arithmetic mean) made in (4).

As disclosed herein, the reference (or initial) value indicated above for the AC region score (e.g., the number of indicator AC regions) may be 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 18, 19, 20 or higher, preferably 5, preferably 8, most preferably 9 or 10, most preferably 10. The reference for the total (e.g., combined) length of the indicator AC regions may be approximately 75, 90, 105, 120, 130, 135, 150, 175, 200, 225, 250, 275, 300, 325, 350, 375, 400, 425, 450. 475, 500 megabases or greater, preferably approximately 75 megabases or greater, preferably approximately 90 or 105 megabases or greater, more preferably approximately 120 or 130 megabases or greater, and more preferably approximately 135 megabases or greater, and most preferably approximately 150 megabases or greater. According to the method of the invention, the reference above for the combined AC region score (e.g., the combined number of indicator LOH regions, indicator TAI regions, and/or indicator LST regions) is 32 or greater and, optionally, may be 34, 36, 38, 40, 42, 44, 46, 48, 50 or greater, preferably 35, preferably 40-44, most preferably > 42. The reference for the total (e.g., combined) length of the indicator LOH regions, indicator TAI regions, and indicator LST regions may be approximately 75, 90, 105, 120, 130, 135, 150, 175, 200, 225, 250, 275, 300, 325, 350, 375. 400, 425, 450, 475, 500 megabases or greater, preferably approximately 75 megabases or greater, preferably approximately 90 or 105 megabases or greater, more preferably approximately 120 or 130 megabases or greater, and more preferably approximately 135 megabases or greater, and most preferably approximately 150 megabases or greater.

In some embodiments, the invention provides a method for detecting an HRD signature in a sample. Therefore, the invention provides a method for detecting an HRD signature in a sample comprising: (1) determining in the sample the total number of LOH regions of a certain size or character (e.g., "indicator LOH regions," as defined herein); (2) determining in the sample the total number of TAI regions of a certain size or character (e.g., "indicator TAI regions," as defined herein); (3) determining in the sample the total number of LST regions of a certain size or character (e.g., "indicator LST regions," as defined herein); (4) combining the determinations made in (1), (2), and (3) (e.g., calculating or deriving a combined AC region score); and (5) characterizing a sample in which the combined AC region score is greater than a reference value having an HRD signature. In some realizations, the reference value may be 42. Therefore, a sample may be characterized as having an HRD signature when the reference value is 42. The reference above for the combined AC region score (i.e., the combined number of indicator LOH regions, indicator TAI regions, and indicator LST regions) may be 32, 34, 36, 38, 40, 42, 44, 46, 48, 50 or higher, preferably 35, preferably 40-44, most preferably >42.

The number of indicator AC regions (or the combined length, an AC region score, or a combined AC region score) in a sample may be considered to be "greater" than a reference if it is at least 2, 3, 4, 5, 6, 7, 8, 9, or 10 times greater than the reference, or it may be considered to be "greater" if it is at least 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater than the reference. In contrast, the number of indicator AC regions (or the combined length, an AC region score, or a combined AC region score) in a sample may be considered "not greater" than a reference if it is no more than 2, 3, 4, 5, 6, 7, 8, 9, or 10 times greater than the reference, or it may be considered "not greater" if it is no more than 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater than the reference.

In some implementations, the reference number (or length, value, or score) is derived from a relevant reference population. Such reference populations may include patients (a) with the same cancer as the patient being examined, (b) with the same cancer subtype, (c) with a cancer that has similar genetic or other clinical or molecular characteristics, (d) who responded to a particular treatment, (e) who did not respond to a particular treatment, (f) who are apparently healthy (e.g., they do not have any cancer or at least do not have the cancer of the patient being examined), etc. The reference number (or length, value, or score) may be (a) representative of the number (or length, value, or score) found in the reference population as a whole, (b) an average (mean, median, etc.) of the number (or length, value, or score) found in the reference population as a whole or in a particular subpopulation, (c) representative of the number (or length, value, or score) (e.g., an average such as mean or median) found in tertiles, quartiles, quintiles, etc., of the reference population ranked by (i) their respective number (or length, value, or score) or (ii) the clinical characteristic they were found to have (e.g., strength of response, prognosis (including time to cancer-specific death), etc.), or (d) selected for having high sensitivity in detecting HRD to predict response to a particular therapy (e.g., platinum, PARP inhibitor, etc.).

In some implementations, if the reference or baseline value is exceeded by the sample test value or score, it indicates that the HRD may be the same as the reference value, which, if not exceeded by the sample test value or score, indicates the absence of HRD (or functional HDR). In some implementations, these values may differ.

A method for predicting the status of the BRCA1 and BRCA2 genes in a sample is described but not claimed herein. This method is analogous to the methods described above and differs in that the determination of AC regions, LOH regions, TAI regions, LST regions, or scores incorporating them, is used to assess (e.g., detect) a BRCA1 and/or BRCA2 deficiency in the sample.

The claimed methods may further comprise a method for predicting a cancer patient's response to a cancer treatment regimen comprising a DNA-damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and/or a PARP inhibitor. Such a method is analogous to the methods described above and differs in that the determination of AC regions, LOH regions, TAI regions, LST regions, or scores incorporating them, including high HRD scores (e.g., a high HRD signature or a high combined AC region score), is used to predict the likelihood that the cancer patient will respond to the cancer treatment regimen.

Patients may be treatment-naïve. For the avoidance of doubt, treatment methods are not included in the claimed invention. A method of cancer treatment is described but not claimed herein. This method is analogous to the methods described above and differs in that a particular treatment regimen (recommended, prescribed, etc.) is administered based, at least in part, on the determination of AC regions, LOH regions, TAI regions, LST regions, or scores incorporating them.

The use of one or more drugs selected from the group consisting of DNA-damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors is described but not claimed herein, in the manufacture of a drug useful for the treatment of cancer in a patient who has been identified as having (or has had) a cancer cell that has been determined to have high levels of HRD (e.g., an HRD signature) as described herein.

This document describes, but does not claim, a method for assessing the presence of a mutation in a gene of an HDR pathway in a sample. This method is analogous to the methods described above and differs in that the determination of AC regions, LOH regions, TAI regions, LST regions, or scores incorporating them, is used to detect (or not) the presence of a mutation in a gene of an HDR pathway.

The invention provides an in vitro method for evaluating the presence of an HDR-deficient state in a patient's cancer cells. The method comprises, or essentially consists of, (a) determining, in a sample comprising the patient's cancer cell, a combined number of indicator LOH regions, indicator TAI regions, and indicator LST regions in at least one pair of human chromosomes from a cancer cell of the patient with cancer, (b) calculating a test value derived from the combined number of indicator LOH regions, indicator TAI regions, and indicator LST regions, and (c) identifying cancer cells with the HDR-deficient state when the test value exceeds a reference value, wherein the reference value is derived from a reference number of 32 or greater. A method for evaluating a patient's cancer cells that have an HRD signature is also disclosed. A method for assessing the presence of a gene mutation in a patient's cancer cells in a gene of an HDR pathway is described but not claimed herein. The method comprises, or essentially consists of, (a) detecting the presence of more than one reference number of AC indicator regions on at least one pair of human chromosomes from a cancer cell of the cancer patient, and (b) identifying that the patient has cancer cells with an HRD signature.

A method for determining whether a patient is likely to respond to a cancer treatment regimen comprising the administration of radiation or a drug selected from the group consisting of DNA-damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors is described but not claimed herein. The method comprises, or essentially consists of, (a) detecting the presence of more than one reference number of AC marker regions on at least one pair of human chromosomes from a cancer cell of the cancer patient, and (b) identifying that the patient is likely to respond to the cancer treatment regimen. A method for evaluating a patient is also described but not claimed herein. The method comprises, or essentially consists of, (a) determining that the patient comprises cancer cells that have an HRD signature, wherein the presence of more than one reference number of AC marker regions on at least one pair of human chromosomes of a cancer cell from the patient with cancer indicates that the cancer cells have the HRD signature, and (b) diagnosing that the patient has cancer cells with the HRD signature. A method for evaluating a patient is described but not claimed herein. The method comprises, or essentially consists of, (a) determining that the patient comprises cancer cells that have an HDR deficiency state, wherein the presence of more than one reference number of AC marker regions on at least one pair of human chromosomes of a cancer cell from the patient with cancer indicates that the cancer cells have the HDR deficiency state, and (b) diagnosing that the patient has cancer cells with the HDR deficiency state. A method for evaluating a patient is described but not claimed herein. The method comprises, or essentially consists of, (a) determining that the patient comprises cancer cells that have an HDR-deficient state, wherein the presence of more than one reference number of AC indicator regions on at least one pair of human chromosomes of a cancer cell from the patient with cancer indicates that the cancer cells have a high HDR, and (b) diagnosing that the patient has cancer cells with an HDR-deficient state. A method for evaluating a patient is described but not claimed herein. The method comprises, or essentially consists of, (a) determining that the patient comprises cancer cells that have a genetic mutation in a gene of an HDR pathway, wherein the presence of more than one reference number of AC indicator regions on at least one pair of human chromosomes of a cancer cell from the patient with cancer indicates that the cancer cells have the genetic mutation, and (b) diagnosing that the patient has cancer cells with the genetic mutation. A method for assessing the likelihood of a patient responding to a cancer treatment regimen comprising the administration of radiation or a drug selected from the group consisting of DNA-damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors is described but not claimed herein. The method comprises, or essentially consists of, (a) determining that the patient comprises cancer cells that have an HRD signature, wherein the presence of more than one reference number of AC marker regions on at least one pair of human chromosomes of a cancer cell of the cancer patient indicates that the cancer cells have the HRD signature, and (b) diagnosing, based at least in part on the presence of the HRD signature, that the patient is likely to respond to the cancer treatment regimen. A method for assessing the likelihood of a patient responding to a cancer treatment regimen comprising the administration of radiation or a drug selected from the group consisting of DNA-damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors is described but not claimed herein. The method comprises, or essentially consists of, (a) determining that the patient comprises cancer cells that have an HRD signature, wherein the presence of more than one reference number of AC marker regions on at least one pair of human chromosomes of a cancer cell of the cancer patient indicates that the cancer cells have an HRD signature, and (b) diagnosing, based at least in part on the presence of the HRD signature, that the patient is likely to respond to the cancer treatment regimen.

A method for performing a diagnostic analysis of a patient's cancer cell is described but not claimed herein. The method comprises, or essentially consists of, (a) detecting the presence of more than one reference number of marker antibody regions on at least one pair of human chromosomes in the cancer cell, and (b) identifying or classifying the patient as having cancer cells with an HRD signature. A method for performing a diagnostic analysis of a patient's cancer cell is described but not claimed herein. The method comprises, or essentially consists of, (a) detecting the presence of more than one reference number of marker antibody regions on at least one pair of human chromosomes in the cancer cell, and (b) identifying or classifying the patient as having cancer cells with an HDR-deficient state. A method for performing a diagnostic analysis of a patient's cancer cell is described but not claimed herein. The method comprises, or essentially consists of, (a) detecting the presence of more than one reference number of AC marker regions on at least one pair of human chromosomes in the cancer cell, and (b) identifying or classifying the patient as having cancer cells with an HDR-deficient status. A method for performing a diagnostic analysis of a cancer cell from a patient is described but not claimed herein. The method comprises, or essentially consists of, (a) detecting the presence of more than one reference number of AC marker regions on at least one pair of human chromosomes in the cancer cell that are longer, and (b) identifying or classifying the patient as having cancer cells with a genetic mutation in a gene of an HDR pathway. A method for performing a diagnostic analysis of a patient's cancer cell to determine whether the cancer patient is likely to respond to a cancer treatment regimen comprising the administration of radiation or a drug selected from the group consisting of DNA-damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors is described but not claimed herein. The method comprises, or essentially consists of, (a) detecting the presence of more than one reference number of AC indicator regions on at least one pair of human chromosomes of the cancer cell, and (b) identifying or classifying the patient as likely to respond to the cancer treatment regimen.

A method for diagnosing that a patient has cancer cells with an HRD signature is described but not claimed herein. The method comprises, or essentially consists of, (a) determining that the patient comprises cancer cells with the HRD signature, wherein the presence of more than one reference number of AC marker regions on at least one pair of human chromosomes of a cancer cell from the patient with cancer indicates that the cancer cells have the HRD signature, and (b) diagnosing that the patient has cancer cells with the HRD signature. A method for diagnosing a patient as having cancer cells with an HDR-deficient state is described but not claimed herein. A method for diagnosing a patient as having cancer cells with an HDR-deficient state is described herein, but not claimed. The method comprises, or essentially consists of, (a) determining that the patient comprises cancer cells that have the HDR-deficient state, wherein the presence of more than one reference number of AC marker regions on at least one pair of human chromosomes of a cancer cell of the patient with cancer indicates that the cancer cells have the HDR-deficient state, and (b) diagnosing that the patient has cancer cells with the HDR-deficient state. A method for diagnosing that a patient has cancer cells with a genetic mutation in a gene of an HDR pathway is described but not claimed herein. The method comprises, or essentially consists of, (a) determining that the patient comprises cancer cells having the genetic mutation, wherein the presence of more than one reference number of AC marker regions on at least one pair of human chromosomes of a cancer cell of the cancer patient indicates that the cancer cells have the genetic mutation, and (b) diagnosing that the patient has cancer cells with the genetic mutation. A method for diagnosing a patient as a candidate for a cancer treatment regimen comprising the administration of radiation or a drug selected from the group consisting of DNA-damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors is described but not claimed herein. The method comprises, or essentially consists of, (a) determining that the patient comprises cancer cells that have an HRD signature, wherein the presence of more than one reference number of AC marker regions on at least one pair of human chromosomes of a cancer cell of the cancer patient indicates that the cancer cells have the HRD signature, and (b) diagnosing, based at least in part on the presence of the HRD signature, that the patient is likely to respond to the cancer treatment regimen. A method for diagnosing a patient as a candidate for a cancer treatment regimen comprising the administration of radiation or a drug selected from the group consisting of DNA-damaging agents, anthracyclines, topoisomerase I inhibitors, and PARP inhibitors is described but not claimed herein. The method comprises, or essentially consists of, (a) determining that the patient comprises cancer cells that have a high HRD signature, wherein the presence of more than one reference number of AC marker regions on at least one pair of human chromosomes of a cancer cell of the cancer patient indicates that the cancer cells have an HRD signature, and (b) diagnosing, based at least in part on the presence of the HRD signature, that the patient is likely to respond to the cancer treatment regimen.

A method for evaluating a patient is described but not claimed herein. The method comprises, or essentially consists of, (a) determining whether the patient has (or has had) cancer cells with more than one reference number of indicator AC regions (or, e.g., an AC region score exceeding a reference AC region score); and (b)(1) diagnosing the patient as having HRD cancer cells if the patient is determined to have (or has had) cancer cells with more than one reference number of AC regions (or, e.g., an AC region score exceeding a reference AC region score); or (b)(2) diagnosing the patient as not having HRD cancer cells if the patient is determined not to have (or has not had) cancer cells with more than one reference number of AC regions (or, e.g., the patient does not have (or has not had) cancer cells with an AC region score exceeding a reference AC region score).

The use of a plurality of oligonucleotides capable of hybridizing with a plurality of polymorphic regions of human genomic DNA is described but not claimed herein, in the manufacture of a diagnostic kit useful for determining the total number or combined length of AC regions in at least one pair of chromosomes (or DNA derived therefrom) in a sample obtained from a cancer patient, and for detecting in the sample (a) HRD, high HRD or probability of HRD (each, e.g., an HDR signature), (b) deficiency (or probability of deficiency) in the sample of a BRCA1 or BRCA2 gene or (c) an increased likelihood of the cancer patient responding to a cancer treatment regimen comprising a DNA-damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation or a PARP inhibitor.

A system for detecting HRD (e.g., an HRD signature) in a sample is described but not claimed herein. The system comprises, or essentially consists of, (a) a sample analyzer configured to produce a plurality of signals on the genomic DNA of at least one pair of human chromosomes (or DNA derived therefrom) in the sample, and (b) a computer subsystem programmed to calculate, based on the plurality of signals, the number or combined length of AC regions on at least one pair of human chromosomes. The computer subsystem may be programmed to compare the number or combined length of AC regions with a reference number to detect in the sample (a) HRD, high HRD, or the probability of HRD (each, e.g., a DRH signature), (b) deficiency (or probability of deficiency) in the sample of a BRCA1 or BRCA2 gene, or (c) an increased probability that the cancer patient will respond to a cancer treatment regimen comprising a DNA-damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, or a PARP inhibitor. The system may include an output module configured to display (a), (b), or (c). The system may also include an output module configured to display a recommendation regarding the use of the cancer treatment regimen.

This document describes, but does not claim, a computer program product embodied in a computer-readable medium that, when executed on a computer, provides instructions for detecting the presence or absence of any AC region along one or more human chromosomes other than the human sex chromosomes X and Y (the AC regions optionally being indicator AC regions); and determining the total number or combined length of AC regions on the one or more pairs of chromosomes. The computer program product may include other instructions.

A diagnostic kit is described but not claimed herein. The kit comprises, or essentially consists of, at least 500 oligonucleotides capable of hybridizing with a plurality of polymorphic regions of human genomic DNA (or DNA derived therefrom); and a software product provided herein. The software product may be incorporated on a computer-readable medium which, when executed on a computer, provides instructions for detecting the presence or absence of any AC region along one or more human chromosomes other than the human sex chromosomes X and Y (the AC regions optionally being indicator AC regions); and determining the total number or combined length of the AC regions on the one or more pairs of chromosomes. The software product may include other instructions.

In some embodiments of the method of the invention described in the preceding paragraphs, any one or more of the following may be applied as appropriate. The AC regions may be determined on at least two, five, ten, or 21 pairs of human chromosomes. The cancer cell may be an ovarian, breast, lung, or esophageal cancer cell. The reference may be 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, or 20 or higher. The at least one pair of human chromosomes may exclude human chromosome 17. The DNA-damaging agent may be cisplatin, carboplatin, oxaliplatin, or picoplatin; the anthracycline may be epirubinic or doxorubicin; the topoisomerase I inhibitor may be campothecin, topotecan, or irinotecan; or the PARP inhibitor may be iniparib, olaparib, or velaparib. The patient may be a patient without prior treatment.

According to the invention, a sample (e.g., a sample of cancer cells or a sample containing DNA derived from one or more cancer cells) can be identified as having an "HRD signature" (or alternatively, an "HDR-deficient signature") if the genome of the cells being evaluated contains a combined AC region score (i.e., the combined number of indicator LOH regions, indicator TAI regions, and indicator LST regions) that exceeds a reference. Conversely, a sample (e.g., a sample of cancer cells or a sample containing DNA derived from one or more cancer cells) can be identified as lacking an "HRD signature" (or alternatively, an "HDR-deficient signature") if the genome of the cells being evaluated contains a combined AC region score that does not exceed a reference.

Cells (e.g., cancer cells) identified as having an HRD signature can be further classified as having a higher probability of HDR deficiency and/or a higher probability of having a deficient state for one or more genes in the HDR pathway. For example, cancer cells identified as having an HRD signature can be further classified as having a higher probability of HDR deficiency. In some cases, cancer cells identified as having an HRD signature can be further classified as having a higher probability of being deficient for one or more genes in the HDR pathway. As used herein, a gene deficiency means that the sequence, structure, expression, and/or activity of the gene or its product is/are deficient compared to normal. Examples include, but are not limited to, low or absent mRNA or protein expression, deleterious mutations, hypermethylation, attenuated activity (e.g., enzyme activity, ability to bind to another biomolecule), etc. As used herein, a deficient pathway status (e.g., of an HDR pathway) means that at least one gene in that pathway (e.g., BRCA1) is deficient. Examples of highly deleterious mutations include frameshift mutations, stop codon mutations, and mutations leading to altered RNA splicing. A deficient status in a gene of the HDR pathway may result in impaired or reduced homology-directed repair activity in cancer cells. Examples of genes in the HDR pathway include, but are not limited to, the genes listed in Table 1.

Table 1. Selected genes from the HDR pathway

As described herein, the identification of loci with AC (as well as the size and number of AC regions) may involve, first, determining the genotype of a sample at several genomic loci (e.g., SNP loci, single bases in large-scale sequencing) and then determining whether each of the loci contains any LOH, TAI, or LST. Any appropriate technique can be used to determine genotypes at loci of interest within a cell's genome. For example, to identify loci that are homozygous or heterozygous, single nucleotide polymorphism (SNP) arrays (e.g., whole-genome human SNP arrays), targeted sequencing of the locus of interest (e.g., sequencing of the SNP locus and its surrounding sequences), and even large-scale sequencing (e.g., whole-exome, transcriptome, or genome sequencing) can be used. Typically, to determine the length of AC regions, an analysis of the homozygous or heterozygous nature of loci along a chromosome can be performed. For example, a stretch of SNP locations that are separated (e.g., approximately 25 kb to 100 kb apart) along a chromosome can be evaluated using the results of the SNP array to determine not only the presence of a homozygous region (e.g., LOH) along a chromosome, but also the length of that region. The results of an SNP array can be used to generate a graph representing allelic dosages along a chromosome. The allelic dosage d<i> of SNP i can be calculated from the adjusted signal intensities of two alleles (A<i> and B<i>): d<i> = A<i/>(A<i> + B<i>). Figures 1 and 2 present an example of such a graph, showing the difference between fresh frozen samples and FFPE, and between SNP microarrays and SNP sequencing analysis. Numerous variations of nucleic acid arrays useful in the invention are known in the technique. These include the arrays used later in the various examples (e.g., the Affymetrix 500K GeneChip array in Example 3; the Affymetrix OncoScan™ FFPE Express 2.0 services (formerly MIP CN Services) in Example 4).

Once the genotype of a sample has been determined with respect to a plurality of loci (e.g., SNPs), standard techniques can be used to identify the loci and regions of LOH, TAI, and LST (including those described in International Application No. PCT/US2011/040953 (published as WO/2011/160063); in International Application No. PCT/US2011/048427 (published as WO/2012/027224); in Popova et al., Ploidy and large-scale genomic instability consistently identify basal-like breast carcinomas with BRCA1/2 inactivation, CANCER RES. (2012) 72:5454-5462). The determination of chromosomal imbalance or large-scale transitions includes determining whether they are somatic or germline aberrations. One way to determine this is by comparing the somatic genotype with the germline genotype. For example, the genotype of a plurality of loci (e.g., SNPs) can be determined in both a germline genotype sample (e.g., blood) and a somatic genotype sample (e.g., tumor). The genotypes of each sample can then be compared (usually using computer techniques) to determine whether the germline cell genome was heterozygous and the somatic cell genome is homozygous. Such loci are LOH loci, and the regions within these loci are LOH regions.

Computational techniques can also be used to determine whether an aberration is germline or somatic. Such techniques are particularly useful when a germline sample is unavailable for analysis and comparison. For example, algorithms such as those described elsewhere can be used to detect LOH regions using SNP array information (Nannya et al., Cancer Res. (2005) 65:6071-6079 (2005)). Typically, these algorithms do not explicitly account for contamination of tumor samples with benign tissue. See International Application No. PCT/US2011/026098 (published as WO 2011/106541) by Abkevich et al.; Goransson et al., PLoS One (2009) 4(6):e6057. This contamination is often high enough to hinder the detection of LOH regions. Improved analytical methods described herein for identifying LOH, TAI, and LST, even in the presence of contamination, include those incorporated into software products such as those described below.

Here is an example. If the observed ratio of signals from two alleles, A and B, is two to one, there are two possibilities. The first possibility is that cancer cells have a locus of hypothalamus (LOH) with deletion of allele B in a sample with 50% contamination from normal cells. The second possibility is that there is no LOH, but allele A is duplicated in a sample without contamination from normal cells. To reconstruct LOH regions based on genotype data (e.g., SNP genotype), an algorithm can be implemented as a computer program like the one described herein. One step of the algorithm is to first reconstruct the allele-specific copy numbers (ASCNs) at each locus (e.g., SNP). The ASCNs are the copy numbers of both paternal and maternal alleles. Then, an LOH region is defined as a segment of the SNPs where one of the ASCNs (paternal or maternal) is equal to zero. The algorithm may be based on maximizing a probability function and may be conceptually similar to a previously described algorithm designed to reconstruct the total copy number (rather than ASCN) at each locus (e.g., SNP). See International Application No. PCT/US2011/026098 (published as WO 2011/106541) by Abkevich et al. The probability function may be maximized over the ASCn of all loci, the level of benign tissue contamination, the genome-wide average total copy number, and the sample-specific noise level. The algorithm's input data may include or consist of (1) sample-specific normalized signal intensities for the two alleles at each locus and (2) a set of assay-specific parameters (specific to different SNP arrays and to the sequence-based approach) defined based on the analysis of a large number of samples with known ASCN profiles.

In some cases, nucleic acid sequencing techniques can be used to genotype loci. For example, genomic DNA can be extracted and fragmented from a cell sample (e.g., from a cancer cell sample). Any appropriate method can be used to extract and fragment genomic nucleic acid, including, but not limited to, commercial kits such as the QlAamp™ DNA Mini Kit (Qiagen™), the MagNA™ Pure DNA Isolation Kit (Roche Applied Science™), and the GenElute™ Mammalian Genomic DNA Miniprep Kit (Sigma-Aldrich™). Once extracted and fragmented, targeted or untargeted sequencing can be performed to determine the genotypes of the sample at the loci. For example, whole-genome, whole-transcriptome, or whole-exome sequencing can be performed to determine genotypes at millions or even billions of base pairs (i.e., base pairs can be "loci" to be evaluated).

In some cases, as an alternative to microarray analysis, targeted sequencing of known polymorphic loci (e.g., SNPs and surrounding sequences) can be performed. For example, genomic DNA can be enriched for fragments containing a locus (e.g., SNP location) to be analyzed using kits designed for this purpose (e.g., Agilent SureSelect™, Illumina TruSeq Capture™, and Nimblegen SeqCap EZ Choice™). Genomic DNA containing the loci to be analyzed can then be hybridized with biotinylated capture RNA fragments to form biotinylated RNA/genomic DNA complexes. Alternatively, DNA capture probes can be used, resulting in the formation of biotinylated DNA/genomic DNA hybrids. Streptavidin-coated magnetic beads and magnetic force can be used to separate biotinylated RNA/genomic DNA complexes from genomic DNA fragments not present in a biotinylated RNA/genomic DNA complex. The resulting biotinylated RNA/genomic DNA complexes can be treated to remove captured RNA from the magnetic beads, leaving unaltered genomic DNA fragments containing a locus to be analyzed. These unaltered genomic DNA fragments containing the loci to be analyzed can be amplified using, for example, PCR techniques. The amplified genomic DNA fragments can then be sequenced using high-throughput sequencing or next-generation sequencing technologies such as Illumina HiSeq™, Illumina MiSeq™, Life Technologies SoLID™, Ion Torrent™, or Roche 454™.

The results of sequencing genomic DNA fragments can be used to identify loci that do or do not contain an allele, analogous to the microarray analysis described herein. In some cases, to determine the length of allele regions, an analysis of the genotype of loci along a chromosome can be performed. For example, a stretch of SNP locations that are separated (e.g., approximately 25 kb to 100 kb apart) along a chromosome can be evaluated by sequencing, and the sequencing results can be used to determine not only the presence of an allele region but also its length. The resulting sequencing data can then be used to generate a graph representing allele dosages along a chromosome. The allelic dose d<i> of SNP i can be calculated from the adjusted number of captured probes of two alleles (A<i> and B<i>): d<i> = A<i>/(A<i> + B<i>). Figures 1 and 2 show an example of such a graph. The determination of whether an aberration is germline or somatic can be performed as described herein.

In some cases, a selection process can be used to select the loci (e.g., SNP loci) to be evaluated using an assay configured to genotype loci (e.g., SNP array-based assays and sequencing-based assays). For example, any human SNP location can be selected for inclusion in an SNP array-based assay or a sequencing-based assay configured to genotype loci. In some cases, 0.5, 1.0, 1.5, 2.0, 2.5 million or more SNP locations present in the human genome may be screened to identify those SNPs that (a) are not present on the Y chromosome, (b) are not mitochondrial SNPs, (c) have a low allele frequency of at least approximately five percent in Caucasians, (d) have a low allele frequency of at least approximately one percent in three non-Caucasian races (e.g., Chinese, Japanese, and Yoruba), and/or (e) do not have a significant deviation from Hardy-Weinberg equilibrium in any of the four races. In some cases, more than 100,000, 150,000, or 200,000 human SNPs meeting criteria (a) through (e) may be selected. From the human SNPs that meet criteria (a) to (e), a group of SNPs (e.g., 110,000 of the major SNPs) can be selected such that the SNPs have a high allele frequency in Caucasians, cover the human genome in a uniformly spaced manner (e.g., at least one SNP every approximately 25 kb to approximately 500 kb), and are not in linkage disequilibrium with any other SNP selected from any of the four races. In some cases, approximately 40,000, 50,000, 60,000, 70,000, 80,000, 90,000, 100,000, 110,000, 120,000, 130,000 or more SNPs that meet each of these criteria can be selected and included in an assay configured to identify AC regions in a human genome. For example, approximately 70,000 to approximately 90,000 (e.g., approximately 80,000) SNPs can be selected for analysis with an SNP array-based assay, and approximately 45,000 to approximately 55,000 (e.g., approximately 54,000) SNPs can be selected for analysis with a sequencing-based assay.

As described herein, any appropriate sample type can be evaluated. For example, a sample containing cancer cells can be evaluated to determine whether the cancer cell genome contains an HRD signature, lacks an HDR signature, has an increased number of indicator AC regions, or has a higher AC region score. Examples of samples containing cancer cells that can be evaluated as described herein include, but are not limited to, tumor biopsy samples (e.g., breast tumor biopsy samples), formalin-fixed and paraffin-embedded tissue samples containing cancer cells, core needle biopsies, fine-needle aspirations, and samples containing cancer cells shed from a tumor (e.g., blood, urine, or other body fluids). For formalin-fixed and paraffin-embedded tissue samples, the sample can be prepared by DNA extraction using an FFPE tissue-optimized genomic DNA extraction kit, including, but not limited to, those described above (e.g., QuickExtract™ FFPE DNA Extraction Kit (Epicentre™) and QIAamp™ DNA FFPE Tissue Kit (Qiagen™)).

In some cases, laser dissection techniques can be performed on a tissue sample to minimize the number of non-cancerous cells in a sample of cancerous cells being evaluated. In some cases, antibody-based purification methods can be used to enrich cancerous cells and/or remove non-cancerous cells. Examples of antibodies that could be used for cancerous cell enrichment include, but are not limited to, anti-EpCAM, anti-TROP-2, anti-c-Met, anti-folate-binding protein, anti-N-Cadherin, anti-CD318, anti-mesenchymal stem cell antigen, anti-Her2, anti-MUC1, anti-EGFR, anti-cytokeratins (e.g., cytokeratin 7, cytokeratin 20, etc.), anti-Caveolin-1, anti-PSA, anti-CA125, and anti-surfactant protein.

Using the methods and materials described herein, any type of cancer cell can be evaluated. For example, breast cancer cells, ovarian cancer cells, liver cancer cells, esophageal cancer cells, lung cancer cells, head and neck cancer cells, prostate cancer cells, colon, rectal or colorectal cancer cells, and pancreatic cancer cells can be evaluated to determine whether the cancer cell genome contains an HRD signature, lacks an HDR signature, has an increased number of AC indicator regions, or has a higher AC region score. The cancer cells can be primary or metastatic cancer cells from ovarian cancer, breast cancer, lung cancer, or esophageal cancer.

When evaluating the genome of cancer cells for the presence or absence of an HRD signature, one or more (e.g., one, two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, or 23) pairs of chromosomes can be assessed.

In some cases, it may be useful to exclude certain chromosomes from this analysis. For example, in women, a pair to be evaluated may include the X sex chromosome pair; whereas in men, a pair of any autosomal chromosome (i.e., any pair other than the X and Y sex chromosome pair) may be evaluated. As another example, in some cases, chromosome pair 17 may be excluded from the analysis. Some chromosomes have been found to carry unusually high levels of AC in certain types of cancer, and therefore, when analyzing samples such as those described herein from patients with these cancers, it may be useful to exclude these chromosomes. In some cases, the sample is from a patient with ovarian cancer, and the chromosome to be excluded is chromosome 17.

Therefore, a predefined number of chromosomes can be analyzed to determine the number of indicator AC regions (or the AC region score or the combined AC region score), preferably the number of AC regions longer than 9 megabases, 10 megabases, 12 megabases, 14 megabases, and most preferably longer than 15 megabases. Alternatively or in addition, the sizes of all identified indicator AC regions can be summed to obtain a total length of indicator AC regions.

As described herein, patients with cancer cells (or samples derived from them) identified as having a high-risk disease response (HRD) signature can be classified, based at least in part on that HRD signature, as likely to respond to a particular cancer treatment regimen. For example, patients with cancer cells exhibiting an HRD signature can be classified, based at least in part on that HRD signature, as likely to respond to a cancer treatment regimen that includes the use of a DNA-damaging agent, a synthetic lethality agent (e.g., a PARP inhibitor), radiation, or a combination thereof. These patients may be treatment-naïve. Examples of DNA-damaging agents include, but are not limited to, platinum-based chemotherapeutic drugs (e.g., cisplatin, carboplatin, oxaliplatin, and picoplatin), anthracyclines (e.g., epirubicin and doxorubicin), topoisomerase I inhibitors (e.g., campothecin, topotecan, and irinotecan), DNA crosslinkers such as mitomycin C, and triazene compounds (e.g., dacarbazine and temozolomide). Synthetic lethality therapeutic approaches generally involve administering an agent that inhibits at least one critical component of a biological pathway that is especially important for the survival of a particular tumor cell. For example, when a tumor cell has a deficient homologous repair pathway (e.g., as determined according to the present invention), poly ADP ribose polymerase inhibitors (or platinum drugs, double-strand break repair inhibitors, etc.) can be especially potent against such tumors because they block two pathways critical for survival (one biologically, e.g., by BRCA1 mutation, and the other synthetically, e.g., by administering a drug of the pathway). Synthetic lethality approaches to cancer therapy are described, e.g., in O'Brien et al., "Converting cancer mutations into therapeutic opportunities," EMBO MOL. MED. (2009) 1:297-299. Examples of synthetic lethality agents include, but are not limited to, PARP inhibitors or double-strand break repair inhibitors in homologous tumor cells deficient in repair, PARP inhibitors in PTEN-deficient tumor cells, methotrexate in MSH2-deficient tumor cells, etc. Examples of PARP inhibitors include, but are not limited to, olaparib, iniparib, and veliparib. Examples of double-strand break repair inhibitors include, but are not limited to, KU55933 (an ATM inhibitor) and NU7441 (a DNA-PKcs inhibitor). Examples of information that may be used in addition to the presence of an HRD signature to base a probability of response to a particular cancer treatment regimen include, but are not limited to, results of previous treatments, mutations in somatic or germline DNA, gene or protein expression profiles (e.g., ER/PR/HER2 status, PSA levels), tumor histology (e.g., adenocarcinoma, squamous cell carcinoma, papillary serous carcinoma, mucinous carcinoma, invasive ductal carcinoma, ductal carcinoma in situ (non-invasive), etc.), disease stage, tumor or cancer grade (e.g., well, moderately, or poorly differentiated (e.g., Gleason classification, modified Bloom-Richardson), etc.), number of previous treatment cycles, etc.

Once classified as likely to respond to a particular cancer treatment regimen (e.g., a cancer treatment regimen that includes the use of a DNA-damaging agent, a PARP inhibitor, radiation, or a combination thereof), the cancer patient may be treated with that cancer treatment regimen. Patients may be treatment-naïve. For the avoidance of doubt, treatment methods are not included in the claimed invention. A method of treating a patient comprising detecting an HRD signature as described herein and administering (or recommending or prescribing) a treatment regimen comprising the use of a DNA-damaging agent, a PARP inhibitor, radiation, or a combination thereof is described but not claimed. Any method appropriate for treating the cancer in question may be used to treat a cancer patient identified with cancer cells that have an HRD signature. For example, platinum-based chemotherapy drugs or a combination of platinum-based chemotherapy drugs, as described elsewhere (see, e.g., U.S. Patents 3,892,790, 3,904,663, 7,759,510, 7,759,488, and 7,754,684), may be used to treat cancer. In some cases, anthracyclines or a combination of anthracyclines may be used to treat cancer, as described elsewhere (see, e.g., U.S. Patents 3,590,028, 4,138,480, 4,950,738, 6,087,340, 7,868,040, and 7,485,707). In some cases, topoisomerase inhibitors may be used. I or a combination of topoisomerase I inhibitors to treat cancer as described elsewhere (see, e.g., U.S. Patents 5,633,016 and 6,403,563). In some cases, PARP inhibitors or a combination of PARP inhibitors may be used to treat cancer as described elsewhere (see, e.g., U.S. Patents 5,177,075, 7,915,280, and 7,351,701). In some cases, radiation may be used to treat cancer as described elsewhere (see, e.g., U.S. Patent 5,295,944). In some cases, a combination comprising different agents may be used to treat cancer (e.g., a combination comprising any of the platinum-based chemotherapeutic drugs, anthracyclines, topoisomerase I inhibitors and/or PARP inhibitors) with or without radiation therapy. In some cases, combination therapy may include any of the above agents or therapies (e.g., a DNA-damaging agent, a PARP inhibitor, radiation, or a combination thereof) together with another agent or therapy, e.g., a taxane agent (e.g., doxetaxel, paclitaxel, abraxane), a growth factor or growth factor receptor inhibitor (e.g., erlotinib, gefitinib, lapatinib, sunitinib, bevacizumab, cetuximab, trastuzumab, panitumumab), and/or an antimetabolite (e.g., 5-fluorouracil, methotrexate).

In some cases, patients identified as having cancer cells lacking a human receptor-deficit (HRD) signature may be classified, based at least in part on a sample lacking an HRD signature, as less likely to respond to a treatment regimen that includes a DNA-damaging agent, a PARP inhibitor, radiation, or a combination thereof. Conversely, such a patient may be classified as likely to respond to a cancer treatment regimen that includes the use of one or more non-HRD-associated cancer treatment agents, such as a taxane agent (e.g., doxetaxel, paclitaxel, abraxane), a growth factor or growth factor receptor inhibitor (e.g., erlotinib, gefitinib, lapatinib, sunitinib, bevacizumab, cetuximab, trastuzumab, panitumumab), and/or an antimetabolite agent (e.g., 5-fluorouracil, methotrexate). Patients may be treatment-naïve. Once classified as likely to respond to a particular cancer treatment regimen (e.g., a cancer treatment regimen that includes the use of a non-HDR-associated cancer treatment agent), the cancer patient may be treated with that cancer treatment regimen. For the avoidance of doubt, treatment methods are not included in the claimed invention. A method of treating a patient is described but not claimed herein, comprising detecting the absence of an HRD signature as described herein and administering (or recommending or prescribing) a treatment regimen that does not include the use of a DNA-damaging agent, a PARP inhibitor, radiation, or a combination thereof. The treatment regimen may include one or more of a taxane agent (e.g., doxetaxel, paclitaxel, abraxane), a growth factor or growth factor receptor inhibitor (e.g., erlotinib, gefitinib, lapatinib, sunitinib, bevacizumab, cetuximab, trastuzumab, panitumumab), and/or an antimetabolite agent (e.g., 5-fluorouracil, methotrexate). Any method appropriate for the cancer being treated may be used to treat a patient with cancer identified as having cancer cells that lack a HRD signature. Examples of information that may be used, in addition to the absence of an HRD signature, to base a classification of the likelihood of response to a particular cancer treatment regimen include, but are not limited to, results of previous treatments, mutations in somatic or germline DNA, gene or protein expression profiles (e.g., ER/PR/HER2 status, PSA levels), tumor histology (e.g., adenocarcinoma, squamous cell carcinoma, papillary serous carcinoma, mucinous carcinoma, invasive ductal carcinoma, ductal carcinoma in situ (non-invasive), etc.), disease stage, tumor or cancer grade (e.g., well, moderately, or poorly differentiated (e.g., Gleason classification, modified Bloom-Richardson), etc.), number of previous treatment cycles, etc.

After a specified period of treatment (e.g., one to six months), the patient can be evaluated to determine whether the treatment regimen is effective. If a beneficial effect is observed, the patient can continue with the same or a similar cancer treatment regimen. If minimal or no beneficial effect is observed, adjustments can be made to the cancer treatment regimen. For example, the dose, frequency of administration, or duration of treatment may be increased. In some cases, other anticancer agents may be added to the treatment regimen, or a particular anticancer agent may be replaced with one or more different anticancer agents. The patient can continue to be monitored as needed, and changes to the cancer treatment regimen can be made as appropriate.

In addition to predicting the likely response to treatment or selecting desirable treatment regimens, an HRD signature can be used to determine a patient's prognosis. A method for determining a patient's prognosis based, at least in part, on detecting the presence or absence of an HRD signature in a patient sample is described but not claimed herein. The method comprises, or essentially consists of, (a) determining whether a patient sample comprises cancer cells (or whether a sample comprises DNA derived from such cells) that has an HRD signature (sometimes referred to herein as having a high HRD) as described herein (e.g., where the presence of more indicator AC regions or an AC region score or combined AC region score is higher than a reference), and (b)(1) determining, based at least in part on the presence of the HRD signature or having a high HRD, that the patient has a relatively good prognosis, or (b)(2) determining, based at least in part on the absence of the HRD signature, that the patient has a relatively poor prognosis. The prognosis may include the patient's probability of survival (e.g., progression-free survival, overall survival), where a relatively good prognosis would include a higher probability of survival compared to a reference population (e.g., the average patient with this type/subtype of cancer, the average patient without an HRD signature, etc.). Conversely, a relatively poor prognosis in terms of survival would include a lower probability of survival compared to a reference population (e.g., the average patient with this type/subtype of cancer, the average patient with an HRD signature, etc.).

As described herein, this document provides methods for evaluating patients for cells (e.g., cancer cells) that have a human liver disease (HLD) signature. One or more medical professionals or practitioners can determine whether a patient sample comprising cancer cells (or if a sample comprises DNA derived from such cells) has an HRD signature. In some cases, one or more medical professionals or practitioners can determine whether a patient contains cancer cells that have an HRD signature by obtaining a cancer cell sample from the patient and evaluating the cancer cell DNA in the sample to determine the presence or absence of an HRD signature as described herein.

In some cases, one or more medical professionals or practitioners may obtain a sample of cancer cells from a patient and provide that sample to a testing laboratory that has the capability to evaluate the cancer cell DNA in the sample to provide an indication of the presence or absence of an HRD signature as described herein. Patients may be those who have not received prior treatment. In such cases, one or more medical professionals or practitioners may determine whether a patient sample comprising cancer cells (or if a sample comprises DNA derived from such cells) has an HRD signature by receiving information about the presence or absence of an HRD signature as described herein directly or indirectly from the testing laboratory. For example, a testing laboratory, after evaluating the DNA of cancer cells for the presence or absence of an HRD signature as described herein, may provide a medical professional or clinician with a written, electronic, or oral report or medical history, or access to them, that provides an indication of the presence or absence of an HRD signature for a particular patient (or patient sample) being evaluated. Such a written, electronic, or oral report or medical history may enable one or more medical professionals or clinicians to determine whether a particular patient being evaluated contains cancer cells that have an HRD signature.

Once a medical professional or team of medical professionals determines that a particular patient under evaluation contains cancer cells with an HRD signature, that professional (or team) may classify that patient as having cancer cells whose genome contains an HRD signature. These patients may be untreated. In some cases, a medical professional or team of medical professionals may diagnose a patient who has been determined to have cancer cells with an HRD signature as having HDR-deficient (or likely HDR-deficient) cancer cells. Such a diagnosis may be based solely on the determination that a sample from the patient comprises cancer cells (or DNA derived from such cells) that have an HRD signature, or it may be based, at least in part, on the determination that a sample from the patient comprises cancer cells (or DNA derived from such cells) that have an HRD signature. For example, a patient who is determined to have cancer cells with an HRD signature may be diagnosed as likely to be HDR deficient based on the combination of the presence of an HRD signature and a deficient status in one or more tumor suppressor genes (e.g., BRCA1/2, RAD51C), a family history of cancer, or the presence of behavioral risk factors (e.g., smoking).

In some cases, a medical professional or team of medical professionals may diagnose a patient who has been determined to have cancer cells with a genome containing an HRD signature as having cancer cells that likely contain genetic mutations in one or more genes of the HDR pathway. These patients may be treatment-naïve. Such a diagnosis may be based solely on the determination that a particular patient being evaluated has cancer cells with a genome containing an HRD signature, or it may be based, at least in part, on the determination that a particular patient being evaluated has cancer cells with a genome containing an HRD signature. For example, a patient who is determined to have cancer cells whose genome contains the presence of an HRD signature may be diagnosed as having cancer cells that likely contain genetic mutations in one or more genes in the HDR pathway based on the combination of the presence of an HRD signature and a family history of cancer or the presence of behavioral risk factors (e.g., smoking).

In some cases, a medical professional or team of medical professionals may diagnose a patient who has been determined to have cancer cells with a human hemorrhagic disease (HRD) signature as having cancer cells likely to respond to a particular cancer treatment regimen. These patients may be treatment-naïve. Such a diagnosis may be based solely on the determination that a sample from the patient comprises cancer cells (or DNA derived from such cells) that have an HRD signature, or it may be based, at least in part, on the determination that a sample from the patient comprises cancer cells (or DNA derived from such cells) that have an HRD signature. For example, a patient who is determined to have cancer cells with an HRD signature may be diagnosed as likely to respond to a particular cancer treatment regimen based on the combination of the presence of an HRD signature and a deficient status in one or more tumor suppressor genes (e.g., BRCA1/2, RAD51), a family history of cancer, or the presence of behavioral risk factors (e.g., smoking). As described herein, a patient determined to have cancer cells with a HRD signature may be diagnosed as likely to respond to a cancer treatment regimen that includes the use of a platinum-based chemotherapy drug such as cisplatin, carboplatin, oxaliplatin, or picoplatin; an anthracycline such as epirubicin or doxorubicin; a topoisomerase I inhibitor such as campothecin, topotecan, or irinotecan; a PARP inhibitor; radiation; a combination thereof; or a combination of any of the above with another anticancer agent. Patients may be treatment-naïve.

Once a medical professional or team of medical professionals determines that a patient sample comprises cancer cells (or if a sample comprises DNA derived from such cells) that have a genome lacking an HRD signature, the medical professional (or team) may classify that patient as having cancer cells whose genome lacks an HRD signature. These patients may be untreated. In some cases, a medical professional or team of medical professionals may diagnose a patient with cancer cells containing a genome lacking an HRD signature as having cancer cells that likely have a functional HDR. In some cases, a medical professional or team of medical professionals may diagnose a patient who has been determined to have cancer cells containing a genome lacking an HRD signature as having cancer cells that likely do not contain genetic mutations in one or more genes in the HDR pathway. In some cases, a medical professional or team of medical professionals may diagnose a patient with cancer cells that contain a genome lacking an HRD signature or containing an increased number of AC regions spanning the entire chromosome. These cancer cells may be less likely to respond to a platinum-based chemotherapy drug such as cisplatin, carboplatin, oxaliplatin, or picoplatin; an anthracycline such as epirubicin or doxorubicin; a topoisomerase I inhibitor such as campothecin, topotecan, or irinotecan; a PARP inhibitor; or radiation. They may also be more likely to respond to a cancer treatment regimen that includes a non-HDR-associated cancer treatment agent such as one or more taxane agents, growth factors or growth factor receptor inhibitors, antimetabolites, etc. These patients may be treatment-naïve.

Methods for performing diagnostic analysis of a nucleic acid sample (e.g., a genomic nucleic acid sample or amplified nucleic acids thereof) from a cancer patient are described but not claimed herein, to determine whether the patient sample comprises cancer cells (or whether the sample comprises DNA derived from such cells) containing an HRD signature and/or an increased number of whole-chromosome AC regions. Patients may be treatment-naïve. For example, one or more laboratory technicians or laboratory professionals may detect the presence or absence of an HRD signature in the genome of the patient's cancer cells (or DNA derived from them) or the presence or absence of an increased number of whole-chromosome AC regions in the genome of the patient's cancer cells. In some cases, one or more laboratory technicians or laboratory professionals may detect the presence or absence of an HRD signature or the presence or absence of an increased number of whole-chromosome AC regions in the patient's cancer cell genome by (a) receiving a sample of cancer cells obtained from the patient, receiving a sample of genomic nucleic acid obtained from cancer cells obtained from the patient, or receiving a sample containing nucleic acids enriched and/or amplified from such genomic nucleic acid sample obtained from cancer cells obtained from the patient and (b) performing an analysis (e.g., an SNP array-based assay or a sequencing-based assay) using the received material to detect the presence or absence of an HRD signature or the presence or absence of an increased number of whole-chromosome AC regions as described herein. In some cases, one or more laboratory technicians or laboratory professionals may receive a sample for analysis (e.g., a sample of cancer cells obtained from the patient, a sample of genomic nucleic acid obtained from cancer cells obtained from the patient, or a sample containing nucleic acids enriched and/or amplified from such a sample of genomic nucleic acid obtained from cancer cells obtained from the patient) directly or indirectly from a medical professional or practitioner. Patients may be treatment-naïve.

Once a laboratory technician or laboratory professional, or group of laboratory technicians or laboratory professionals, detects the presence of an HRD signature as described herein, the laboratory technician or laboratory professional (or group) may associate that HRD signature, or the result (or results, or a summary of results) of the diagnostic test performed, with the corresponding patient's name, medical history, symbolic/numeric identifier, or a combination thereof. Such identification may be based solely on the detection of the presence of an HRD signature or may be based, at least in part, on the detection of the presence of an HRD signature. For example, a laboratory technician or lab professional may identify a patient with cancer cells detected as having an HRD signature as possibly having HDR-deficient cancer cells (or being more likely to respond to a particular treatment, as described in detail throughout this document) based on a combination of the presence of an HRD signature and the results of other genetic and biochemical tests performed in the testing laboratory. These patients may be treatment-naïve.

The opposite is also true. Specifically, once a laboratory technician or laboratory professional, or group of laboratory technicians or laboratory professionals, detects the absence of an HRD signature, the laboratory technician or laboratory professional (or group) may associate the absence of an HRD signature, or the result (or results, or a summary of results) of the diagnostic analysis performed, with the corresponding patient's name, medical history, symbolic/numeric identifier, or a combination thereof. In some cases, a laboratory technician or laboratory professional, or group of laboratory technicians or laboratory professionals, may identify a patient with cancer cells detected as lacking an HRD signature as having cancer cells with possibly unaltered HDR (or as having a lower likelihood of responding to a particular treatment as described throughout this document), either based solely on the absence of an HRD signature or on a combination of the presence of an HRD signature and the results of other genetic and biochemical tests performed in the testing laboratory. Patients may be patients who have not received prior treatment.

Generally, the results of any analysis according to the invention will be communicated to physicians, genetic counselors, and/or patients (or other interested parties, such as researchers) in a transmissible form that can be communicated or transmitted to any of the aforementioned parties. This form may vary and may be tangible or intangible. The results may be incorporated into descriptive statements, diagrams, photographs, tables, images, or any other visual form. For example, charts or diagrams displaying information about the genotype or LOH (or HRD status) may be used to explain the results. The statements and visual forms may be recorded on a tangible medium such as paper, computer-readable media such as a floppy disk, compact disc, flash memory, etc., or on an intangible medium, e.g., an electronic medium in the form of email or a website on the Internet or intranet. In addition, the results may also be recorded in the form of sound and transmitted through any suitable medium, e.g. e.g., via analog or digital cable lines, fiber optic cables, etc., by telephone, fax, wireless mobile phone, internet phone and similar means.

Therefore, test result information and data can be generated anywhere in the world and transmitted to a different location. As an illustrative example, when a trial is conducted outside the United States, test result information and data can be generated, issued in a transmissible form as described above, and then imported into the United States. Accordingly, a method for producing a transmissible form of information about an HRD signature for at least one patient sample may comprise the steps of (1) determining an HRD signature according to the methods described herein; and (2) incorporating the result of the determination step into a transmissible form. The transmissible form is a product of such a method.

Several methods described herein involve a step of correlating the presence of an HRD signature as described herein (e.g., the total number of indicator AC regions or an AC region score or a combined AC region score greater than a reference) with a particular clinical feature (e.g., an increased likelihood of a BRCA1 or BRCA2 gene deficiency; an increased likelihood of HDR deficiency; an increased likelihood of response to a treatment regimen comprising a DNA-damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and/or a PARP inhibitor; etc.) and, optionally, correlating the absence of an HRD signature with one or more distinct clinical features. Throughout this document, another method may involve, in addition to or instead of a correlation step, one or both of the following steps: (a) concluding that the patient has the clinical feature based, at least in part, on the presence or absence of the HRD signature; or (b) communicate that the patient has the clinical characteristic based, at least in part, on the presence or absence of the HRD signature.

By way of illustration, a method for predicting a cancer patient's response to a cancer treatment regimen comprising a DNA-damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and/or a PARP inhibitor is described herein, the method comprising: (1) determining in a sample (a) a LOH region score for the sample; (b) a TAI region score for the sample; (c) an LST region score for the sample; and (2)(a) correlating a combination of the LOH region score, the TAI region score, and the LST region score (e.g., a combined AC region score) that exceeds a reference, with a higher likelihood of responding to the treatment regimen; or optionally (2)(b) correlating a combination of the LOH region score, the TAI region score, and the LST region score (e.g., a combined AC region score) that does not exceed a reference, with a probability no greater than responding to the treatment regimen; or optionally (2)(c) correlating an average (e.g., arithmetic mean) of the LOH region score, the TAI region score, and the LST region score. Pursuant to the preceding paragraph, this description is understood to include a description of two alternative related methods. One is a method for predicting the response of a cancer patient to a cancer treatment regimen comprising a DNA-damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and/or a PARP inhibitor, said method comprising: (1) determining in a sample (a) an LOH region score for the sample; (b) a TAI region score for the sample; (c) an LST region score for the sample; or (d) an average (e.g., arithmetic mean) of the LOH region score, the TAI region score, and the LST region score; and (2)(a) conclude that such patient has a higher likelihood of responding to such cancer treatment regimen based, at least in part, on a combination of the LOH region score, the TAI region score, and the LST region score (e.g., a combined AC region score) that exceeds a reference; or optionally (2)(b) conclude that such patient is not more likely to respond to said cancer treatment regimen based, at least in part, on a combination of the LOH region score, the TAI region score, and the LST region score (e.g., a combined AC region score), or an average (e.g., arithmetic mean) of the LOH region score, the TAI region score, and the LST region score, not exceeding a reference value. Another is a method for predicting the response of a cancer patient to a cancer treatment regimen comprising a DNA-damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, and/or a PARP inhibitor, said method comprising: (1) determining in a sample (a) an LOH region score for the sample; (b) a TAI region score for the sample; (c) an LST region score for the sample; or (d) an average (e.g., arithmetic mean) of the LOH region score, the TAI region score, and the LST region score; and (2)(a) communicate that such patient has a higher likelihood of responding to such cancer treatment regimen based, at least in part, on a combination of the LOH region score, the TAI region score, and the LST region score (e.g., a combined AC region score); or an average (e.g., arithmetic mean) of the LOH region score, the TAI region score, and the LST region score, which exceeds a reference; or optionally (2)(b) communicate that such patient is not more likely to respond to such cancer treatment regimen based, at least in part, on a combination of the LOH region score, the TAI region score and the LST region score (e.g., a combined AC region score); or an average (e.g., arithmetic mean) of the LOH region score, the TAI region score and the LST region score, which does not exceed a reference.

In each case described in this document, which involves correlating a particular trial or analytical result (e.g., total number of indicator AC regions greater than a reference number, presence of an HRD signature, etc.) with some probability (e.g., increase, no increase, decrease, etc.) of some clinical characteristic (e.g., response to a particular treatment, specific cancer death, etc.), or additionally or alternatively concluding or communicating such clinical characteristic based, at least in part, on such particular trial or analytical result, such conclusive or communicative correlation may include assigning a risk or probability of the clinical characteristic occurring based, at least in part, on the result of the particular trial or analytical result. Such risk is a percentage probability of the event or outcome occurring. The patient may be assigned to a risk group (e.g., low risk, intermediate risk, high risk, etc.). "Low risk" is any probability percentage below 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, or 50%. "Medium risk" is any probability percentage above 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, or 50% and below 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, or 75%. "High risk" is any probability percentage above 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, or 99%.

As used herein, "communicating" certain information means making that information known to another person or transferring that information to something (e.g., a computer). A patient's prognosis or the likelihood of their response to a particular treatment may be communicated. The information used to arrive at that prognosis or response prediction may also be communicated (e.g., HRD signature in accordance with the present invention, etc.). This communication may be auditory (e.g., verbal), visual (e.g., written), electronic (e.g., data transferred from one computer system to another), etc. Communicating a cancer classification (e.g., prognosis, likelihood of response, appropriate treatment, etc.) may involve generating a report that communicates the cancer classification. The report may be a paper report, an audio report, or an electronic report. The report may be displayed and/or stored on a computing device (e.g., a portable device, a desktop computer, a smart device, a website, etc.). The cancer classification can be communicated to a physician (e.g., the physician is provided with a report communicating the classification). The cancer classification can also be communicated to a patient (e.g., the patient is provided with a report communicating the classification). Communication of a cancer classification can also be accomplished by transferring information (e.g., data) containing the classification to a server computer and allowing an intermediary or end user to access that information (e.g., view the information as displayed on the server, download the information as one or more files transferred from the server to the intermediary's or end user's device, etc.).

Whenever any aspect of the disclosure involves the conclusion of a fact (e.g., a patient's prognosis or the likelihood of a patient's response to a particular treatment regimen), this may include a computer program that concludes such a fact, typically after performing an algorithm that applies information about AC regions in accordance with this disclosure.

In each of the embodiments described herein that involves a number of indicator AC regions or an average (e.g., arithmetic mean) of the combined AC region scores, the present invention encompasses a related embodiment that involves a test value or score derived from such number. In other words, it is not necessary to use the indicator AC region numbers discovered in the various methods and systems described herein; a test value or score derived from such numbers may be used. For example, a method of treating cancer in a patient is described herein but not claimed, comprising: (1) determining in a sample from said patient two or more of, or an average (e.g., arithmetic mean) of, (a) the number of indicator LOH regions, (b) the number of indicator TAI regions, or (c) the number of indicator LST regions; (2) provide one or more test values derived from such number of indicator LOH regions, indicator TAI regions and/or indicator LST regions; (3) compare such test value(s) with one or more reference values (e.g., reference values derived from the number of indicator LOH regions, indicator TAI regions and/or indicator LST regions in a reference population (e.g., mean, median, tertiles, quartiles, quintiles, etc.)); and (4)(a) administer to such patient a cancer drug, or recommend or prescribe or initiate a treatment regimen comprising chemotherapy and/or a synthetic lethality agent based, at least in part, on such comparator stage revealing that one or more of the test values is greater (e.g., at least 2, 3, 4, 5, 6, 7, 8, 9 or 10 times greater; at least 1, 2, 3, 4, 5, 6, 7, 8, 9 or 10 standard deviations greater) than at least one of such reference values; or optionally (4)(b) recommend, prescribe, or initiate a treatment regimen that does not include chemotherapy and/or a synthetic lethal agent based at least in part on such a comparison stage that reveals that one or more of the test values is no greater (e.g., no more than 2, 3, 4, 5, 6, 7, 8, 9, or 10 times greater; no more than 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 standard deviations greater) than at least one of such reference values. For the avoidance of doubt, treatment methods are not included in the claimed invention. The same applies where the test value or score is used to determine the patient's prognosis, the likelihood of the patient's response to a particular treatment regimen, the likelihood of the patient or the patient's sample having a BRCA1, BRCA2, RAD51C, or HDR deficiency, etc.

Figure 8 shows an illustrative process by which a computer system (or a computer program (e.g., software) containing computer-executable instructions) can identify LOH loci or regions from genotype data as described herein. As will be obvious to those skilled in the art, this process can be adapted for use in the determination of TAI and LST. If the observed ratio of signals from two alleles, A and B, is two to one, there are two possibilities. The first possibility is that cancer cells have LOH with deletion of allele B in a sample with 50% contamination from normal cells. The second possibility is that there is no LOH, but allele A is duplicated in a sample without contamination from normal cells. The process begins at box 1500, where the computer system collects the following data; (1) sample-specific normalized signal intensities of both alleles at each locus and (2) a set of assay-specific parameters (specific to different SNP arrays and to the sequence-based approach) defined based on the analysis of a large number of samples with known ASCN profiles. As described herein, any appropriate assay, such as an SNP array-based assay or a sequencing-based assay, may be used to assess locus homozygosity or heterozygosity along a chromosome. In some cases, a system including a signal sensor and a computer may be used to collect data (e.g., fluorescent signals or sequencing results) regarding the homozygous or heterozygous nature of the locus plurality (e.g., sample-specific normalized signal intensities at both alleles at each locus). In box 1510, the allele-specific copy numbers (ASCNs) (e.g., for each SNP) are reconstructed for each allele. ASCNs are the copy numbers of both paternal and maternal alleles. In box 1530, a probability function is used to determine whether a homozygous locus or a homozygous locus region is due to a locus of homozygosity (LOH). This can be conceptually analogous to an algorithm described earlier designed to reconstruct the total number of copies (rather than ASCNs) at each locus (e.g., SNPs). See International Application No. PCT/US2011/026098 (published as WO 2011/106541) by Abkevich et al. The probability function can be maximized over the ASCN of all loci, the level of contamination with benign tissue, the total genome-averaged copy number, and the sample-specific noise level. In box 1540, a region of LOH is defined as a stretch of SNPs where one of the ASCNs (paternal or maternal) is zero. The computer process may also include a stage of investigating or determining whether a patient has not received prior treatment.

Figure 3 shows an illustrative process by which a computer system can determine the presence or absence of a LOH signature and is included to illustrate how this process can, as will be obvious to someone skilled in the art, be applied to TAI and LST. The process begins at box 300, where the computer system collects data regarding the homozygous or heterozygous nature of a plurality of loci along a chromosome. As described herein, to assess locus homozygosity or heterozygosity along a chromosome, any appropriate assay can be used, such as an SNP array-based assay or a sequencing-based assay. In some cases, a system that includes a signal sensor and a computer can be used to collect data (e.g., fluorescent signals or sequencing results) regarding the homozygous or heterozygous nature of the plurality of loci. In box 310, data relating to the homozygous or heterozygous nature of a plurality of loci, as well as the location or spatial relationship of each locus, are evaluated by the computer system to determine the length of any of the LOH regions present along a chromosome. In box 320, the computer system evaluates data relating to the number of LOH regions detected and the length of each detected LOH region to determine the number of LOH regions that have a length (a) greater than or equal to a predetermined number of Mb (e.g., 15 Mb) and (b) less than the entire length of the chromosome containing that LOH region. Alternatively, the computer system can determine the total or combined LOH length as described above. In box 330, the computer system formats a result that provides an indication of the presence or absence of an HRD signature. Once formatted, the computer system can present the result to a user (e.g., a laboratory technician, medical professional, or clinician). As described herein, the presence or absence of an HRD signature can be used to provide an indication of a patient's likely HDR status, an indication of the likely presence or absence of genetic mutations in genes of the HDR pathway, and/or an indication of potential cancer treatment regimens.

Figure 4 is a diagram of an example of a 1400 computing device and a 1450 portable computing device, which can be used with the techniques described herein. The 1400 computing device is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The 1450 computing device is intended to represent various forms of portable devices, such as personal digital assistants, mobile phones, smartphones, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are intended to be illustrative only.

The 1400 computer device includes a 1402 processor, a 1404 memory, a 1406 storage device, a 1408 high-speed interface connecting to the 1404 memory and 1410 high-speed expansion ports, and a 1415 low-speed interface connecting to a 1414 low-speed bus and the 1406 storage device. Each of the components 1402, 1404, 1406, 1408, 1410, and 1415 are interconnected by various buses and can be mounted on a common motherboard or in other ways, as appropriate. The 1402 processor can process instructions for execution within the 1400 computing device, including instructions stored in memory 1404 or storage device 1406 to display graphical information for a GUI (graphical user interface) on an external input/output device, such as a 1416 display connected to the 1408 high-speed interface. In other implementations, multiple processors and/or multiple buses can be used, as appropriate, along with multiple memories and memory types. Multiple 1400 computing devices can also be connected, with each device providing parts of the necessary operations (e.g., as a server bank, a group of blade servers, or a multiprocessor system).

Memory 1404 stores information within the 1400 computer device. In one implementation, memory 1404 is a volatile memory unit or units. In another implementation, memory 1404 is a non-volatile memory unit or units. Memory 1404 can also be another form of computer-readable media, such as a magnetic or optical disk.

The 1406 storage device is capable of providing mass storage for the 1400 computing device. In one implementation, the 1406 storage device may be or contain a computer-readable medium, such as a floppy disk device, a hard disk device, an optical disk device, or a tape device, flash memory or other similar solid-state memory device, or an array of devices, including devices in a storage area network or other configurations.

A computer program product can be tangibly embodied on a data carrier. The computer program product can also contain instructions that, when executed, perform one or more methods, such as those described herein. The data carrier is a computer-readable medium, such as memory (1404), storage device (1406), memory in the processor (1402), or a propagated signal.

The high-speed controller 1408 handles bandwidth-intensive operations for the computing device 1400, while the low-speed controller 1415 handles lower-bandwidth-intensive operations. This role assignment is for illustrative purposes only. In one implementation, the high-speed controller 1408 is coupled to memory 1404, the display 1416 (e.g., via a graphics processor or accelerator), and the high-speed expansion ports 1410, which can accept various expansion cards (not shown). In the implementation, the low-speed controller 1415 is coupled to the storage device 1406 and the low-speed expansion port 1414. The low-speed expansion port, which can include various communication ports (e.g., USB, Bluetooth, Ethernet, or wireless Ethernet), can be coupled to one or more input/output devices, such as a keyboard, pointing device, scanner, optical reader, fluorescent signal detector, or a network device such as a switch or router, e.g., via a network adapter.

The 1400 computing device can be implemented in several different ways, as shown in the figure. For example, it can be implemented as a standard 1420 server, or multiple times in a group of such servers. It can also be implemented as part of a 1424 rack server system. Furthermore, it can be implemented in a personal computer such as a 1422 laptop. Alternatively, the components of the 1400 computing device can be combined with other components in a portable device (not shown), such as the 1450 device. Each of these devices can contain one or more 1400 and 1450 computing devices, and a complete system can consist of multiple 1400 and 1450 computing devices communicating with each other.

The 1450 computer device includes a 1452 processor, 1464 memory, an input/output device such as a display 1454, a 1466 communication interface, and a 1468 transceiver, among other components (e.g., a scanner, an optical reader, a fluorescent signal detector). The 1450 device may also be equipped with a storage device, such as a microdrive or other device, to provide additional storage. Each of the components 1450, 1452, 1464, 1454, 1466, and 1468 is interconnected by various buses, and several of the components may be mounted on a common motherboard or in other ways, as appropriate.

The 1452 processor can execute instructions within the 1450 computing device, including instructions stored in the 1464 memory. The processor can be implemented as a chipset comprising multiple separate analog and digital processors. The processor can, for example, coordinate the other components of the 1450 device, such as controlling user interfaces, applications run by the 1450 device, and wireless communication.

The processor 1452 can communicate with a user through the control interface 1458 and the display interface 1456 coupled to the display 1454. The display 1454 can be, for example, a TFT LCD (thin-film transistor liquid crystal display) or an OLED (organic light-emitting diode) display, or another suitable display technology. The display interface 1456 can comprise appropriate circuitry to activate the display 1454 to present graphics and other data to a user. The control interface 1458 can receive commands from a user and convert them for transmission to the processor 1452. Additionally, an external interface 1462 can be provided in communication with the processor 1452 to enable near-area communication between the device 1450 and other devices. The external interface 1462 can provide, for example, wired communication in some implementations, or wireless communication in others, and multiple interfaces can also be used.

Memory 1464 stores information within the computing device 1450. Memory 1464 can be implemented as one or more computer-readable media, a volatile memory unit or units, or a non-volatile memory unit or units. Expansion memory 1474 can also be provided and connected to the device 1450 via the expansion interface 1472, which may include, for example, a SIMM (Simple In-line Memory Module) card interface. Such expansion memory 1474 can provide additional storage space for the device 1450, or it can store applications or other data for the device 1450. For example, expansion memory 1474 can include instructions to carry out or supplement the processes described above, and it can also include secure information. Therefore, for example, the 1474 expansion memory can be provided as a security module for the 1450 device and can be programmed with instructions that allow secure use of the 1450 device. In addition, secure applications can be provided via SIMM cards, along with additional information, such as placing identification information on the SIMM card in a way that prevents it from being hacked.

Memory may include, for example, flash memory and/or NVRAM, as explained later. In an implementation, a software product is tangibly embodied in a storage medium. The software product contains instructions that, when executed, perform one or more methods, such as those described herein. The storage medium is a machine-readable medium, such as memory (1464), expansion memory (1474), memory in the processor (1452), or a propagated signal that can be received, for example, on transceiver (1468) or external interface (1462).

The 1450 device can communicate wirelessly via the 1466 communication interface, which may include digital signal processing circuitry where required. The 1466 communication interface can provide communication in various modes or protocols, such as GSM voice calls, SMS, Em S or MMS messaging, CDMA, TDMA, PDC, WCDMA, CDMA2000, or GPRS, among others. Such communication can occur, for example, via the 1468 radio frequency transceiver. Short-range communication, such as via Bluetooth, Wi-Fi, or another similar transceiver (not shown), is also possible. Furthermore, the 1470 GPS (Global Positioning System) receiver module can provide the 1450 device with additional wireless navigation and location data, which applications running on the 1450 device can use as appropriate.

The 1450 device can also communicate audibly using an audio codec 1460, which can receive spoken information from a user and convert it into usable digital information. The audio codec 1460 can also generate audible sound for a user, such as through a speaker, e.g., in a headset of the 1450 device. This sound can include the sound of voice telephone calls, recorded sound (e.g., voicemails, music files, etc.), and sound generated by applications running on the 1450 device.

The 1450 computing device can be implemented in several different forms, as shown in the figure. For example, it can be implemented as a 1480 mobile phone. It can also be implemented as part of a 1482 smartphone, personal digital assistant, or other similar mobile device.

Various implementations of the systems and techniques described in this document can be carried out in digital electronic circuits, integrated circuits, ASICs (application-specific integrated circuits), computer equipment, microprograms, computer programs, and/or combinations thereof. These various implementations may include their implementation in one or more computer programs that can be executed and/or interpreted on a programmable system that includes at least one programmable processor, which may be special-purpose or general-purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

These computer programs (also known as programs, software, applications, or software code) include computer instructions for a programmable processor and may be implemented in a high-level procedural and/or object-oriented programming language and/or in assembly/computer language. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to a computer program product, apparatus, and/or device (e.g., magnetic disks, optical disks, memory, and programmable logic devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide computer instructions and/or data to a programmable processor.

To facilitate user interaction, the systems and techniques described herein can be implemented on a computer that has a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) to show information to the user and a keyboard and pointing device (e.g., a mouse or scroll ball) through which the user can input data into the computer. Other types of devices can also be used to enable user interaction; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual, auditory, or tactile feedback); and user input can be received in any form, including acoustic, voice, or tactile input.

The systems and techniques described herein can be implemented in a computer system that includes a back-end component (e.g., a data server), a middleware component (e.g., an application server), a front-end component (e.g., a client computer with a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described herein), or any combination of such back-end, middleware, and front-end components. The system components can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

A computer system can include clients and servers. A client and server are usually remote from each other and typically interact through a communication network. The relationship between the client and the server arises from the computer programs that run on their respective computers and that have a client-server relationship with each other.

In some cases, a computer system provided herein may be configured to include one or more sample analyzers. A sample analyzer may be configured to produce multiple signals on the genomic DNA of at least one pair of human chromosomes from a cancer cell. For example, a sample analyzer may produce signals that can be interpreted in a way that identifies the genotype of loci along a chromosome. In some cases, a sample analyzer may be configured to perform one or more steps of an SNP array-based assay or a sequencing-based assay and may be configured to produce and/or capture signals from such assays. In some cases, a computer system provided herein may be configured to include a computing device. In such cases, the computing device may be configured to receive signals from a sample analyzer. The computing device may include computer-executable instructions or a computer program (e.g., software) containing computer-executable instructions to perform one or more of the methods or steps described herein. In some cases, such computer-executable instructions may instruct a computer device to analyze signals from a sample analyzer, from another computer device, from an SNP array-based assay, or from a sequencing-based assay. The analysis of such signals can be carried out to determine genotypes, homozygosity or other chromosomal aberrations at certain loci, AC regions, the number of AC regions, to determine the size of AC regions, to determine the number of AC regions that have a particular size or size range, to determine whether a sample is positive or not for an HRD signature, to determine the number of Indicator AC regions on at least one pair of human chromosomes, to determine the probability of a deficiency in the BRCA1 and/or BRCA2 genes, to determine the probability of an HDR deficiency, to determine the probability that a cancer patient will respond to a particular cancer treatment regimen (e.g., a regimen that includes a DNA-damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, a PARp inhibitor, or a combination thereof), or to determine a combination of these elements.

In some cases, a computer system described herein may include computer-executable instructions or a computer program (e.g., software) containing computer-executable instructions for formatting a result that provides an indication of the number of AC regions, the size of AC regions, the number of AC regions that have a particular size or size range, whether a sample is positive or not for an HRD signature, the number of indicator AC regions on at least one pair of human chromosomes, a probability of a deficiency in the BRCA1 and/or BRCA2 genes, for determining the probability of an HDR deficiency, the probability that a cancer patient will respond to a particular cancer treatment regimen (e.g., a regimen that includes a DNA-damaging agent, an anthracycline, a topoisomerase I inhibitor, radiation, a PARp inhibitor, or a combination thereof), or a combination of these elements. In some cases, a computer system described herein may include computer-executable instructions or a computer program (e.g., software) containing computer-executable instructions to determine a desired cancer treatment regimen for a particular patient based, at least in part, on the presence or absence of an HRD signature or the number of indicator AC regions.

In some cases, a computer system described herein may include a preprocessing device configured to process a sample (e.g., cancer cells) so that an SNP array-based assay or a sequencing-based assay can be performed. Examples of preprocessing devices include, but are not limited to, devices configured to enrich cell populations of cancer cells rather than non-cancer cells, devices configured to cause cell lysis and/or extract genomic nucleic acid, and devices configured to enrich a sample with particular genomic DNA fragments.

This document also describes kits for evaluating samples (e.g., cancer cells) as described herein. For example, this document discloses kits for evaluating the presence of an HRD signature in cancer cells or for determining the number of AC marker regions on at least one pair of human chromosomes. A kit described herein may include SNP probes (e.g., an array of SNP probes for performing an SNP array-based assay described herein) or primers (e.g., primers designed to sequence SNP regions through a sequencing-based assay) along with a software product containing computer-executable instructions for carrying out one or more of the methods or steps described herein (e.g., computer-executable instructions for determining the number of A<sup>c> marker regions). In some cases, a kit described herein may include at least 500, 1,000, 10,000, 25,000, or 50,000 SNP probes capable of hybridizing to polymorphic regions of human genomic DNA. In some cases, a kit described herein may include at least 500, 1,000, 10,000, 25,000, or 50,000 primers capable of sequencing polymorphic regions of human genomic DNA. In some cases, a kit described herein may include one or more distinct ingredients for performing an SNP array-based assay or a sequencing-based assay. Examples of such distinct ingredients include, but are not limited to, buffers, sequencing nucleotides, enzymes (e.g., polymerases), etc. This document also describes the use of any appropriate number of the materials described herein in the fabrication of a kit to carry out one or more of the methods or steps described herein. For example, this document describes the use of a collection of SNP probes (e.g., a collection of 10,000 to 100,000 SNP probes) and a software product described herein in the fabrication of a kit to assess the presence of an HRD signature in cancer cells. As another example, this document describes the use of a collection of primers (e.g., a collection of 10,000 to 100,000 primers for sequencing SNP regions) and a software product described herein in the fabrication of a kit to assess the presence of an HRD signature in cancer cells.

The following describes illustrative but not limiting details of methods and systems in accordance with the more general description above.

The sample used may be a frozen tumor sample. The sample may be from a particular breast cancer subtype chosen from triple-negative, ER+/HER2-, ER-/HER2+, or ER+/HER2+. The laboratory assay portion of the method, system, etc., may comprise analyzing the sample to sequence the BRCA1 and/or BRCA2 genes (as well as any other gene or genes from Table 1). The laboratory assay portion of the method, system, etc., may comprise analyzing the sample to determine the allelic dosage (e.g., genotype, copy number, etc.) for at least 10,000, 20,000, 30,000, 40,000, 50,000, 60,000, 70,000, 80,000, 90,000, 100,000, or more selected genome-wide SNPs. SNP analysis can be performed using an oligonucleotide microarray as described above. BRCA sequence analysis, SNP analysis, or both can be performed using a capture probe (e.g., probes for each SNP to be analyzed and/or probes to capture the entire coding region of BRCA1 and/or BRCA2) with a subsequent PCR enrichment technique (e.g., Agilent™ SureSelect XT). BRCA sequence analysis, SNP analysis, or both can be performed by processing the enrichment result using a next-generation sequencing platform (e.g., Illumina™ HiSeq2500). The sample can be analyzed for somatic and/or germline mutations of BRCA1/2, which may include large rearrangements. The sample can be analyzed for BRCA1 promoter methylation (e.g., by a cPCR (quantitative PCR) assay (e.g., SA Biosciences)). A sample can be determined to have high methylation (or be "methylated") if the sample has a methylation level greater than 10% (or 5%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%) (e.g., % of methylated CpGs in the BRCA1 or BRCA2 promoter). DNA from compatible (non-tumor) normal tissue from a patient can be analyzed, e.g., to determine whether BRCA1 or BRCA2 mutations are germline or somatic.

In some embodiments, the LOH region score can be calculated by counting the number of LOH regions that are >15 Mb in length but shorter than the length of a complete chromosome. In some embodiments, the TAI region score can be calculated by counting the number of allelic imbalance telomeric regions that are >11 Mb in length, extending to one of the subtelomeres but not across the centromere. In some embodiments, the LST region score can be calculated by counting the number of breakpoints between regions longer than 10 megabases that have a stable number of regions after filtering out regions shorter than 3 megabases. The LST region score can be modified by adjusting for ploidy: LSTm = LST - kP, where P is ploidy and k is a constant (in some embodiments, k = 15.5). BRCA1/2 deficiency can be defined as loss of function resulting from a BRCA1 or BRCA2 mutation, or methylation of the BRCA1 or BRCA2 promoter region, along with LOH in the affected gene. The response to treatment may be a partial complete response ("pCR"), which in some realizations may be defined as Miller-Payne stage 5 after treatment (e.g., neoadjuvant).

BRCA deficiency can be predicted with a p-value (probability) of at least 8*10-12, 6*10-6, 0.0009, 0.01, 0.03, 2*10-16, 3*10-6, 10-6, 0.0009, 8*10-12, 2*10-16, 8*10-8, 6*10-6, 3*10-6, or 0.0002 (i.e., each AC region score is predefined, and multiple scores are optionally combined to yield these p-values). The p-values can be calculated using the Kolmogorov-Smirnov test. HRD scores and age at diagnosis can be coded as a numerical variable (e.g., whole number), breast cancer stage and subtype can be coded as categorical variables, and grade can be analyzed as a numerical or categorical variable, or both.

The p-values may be two-sided. Logistic regression analysis can be used to predict BRCA1/2 deficiency based on an HRD score as disclosed herein (including the combined HRD score). The various AC region scores can be correlated according to (e.g., defined to achieve) the following correlation coefficients: LOH region score and TAI region score = 0.69 (p = 10-39), between LOH and LST = 0.55 (p = 2*10-19), and between TAI and LST = 0.39 (p = 10-9).

According to the methods disclosed herein, the LOH region score, TAI region score, and LST region score can be combined as follows to detect BRCA1/2 deficiency and/or predict response to therapy (e.g., response to platinum therapy, e.g., cisplatin): Combined AC region score = 0.21*LOH region score, 0.67*TAI region score, 0.12*LST region score. Alternatively, the LOH region score, TAI region score, and LST region score can be combined as follows to detect BRCA1/2 deficiency and/or predict response to therapy (e.g., response to platinum therapy, e.g., cisplatin): Combined AC region score = 0.11*LOH region score 0.25*TAI region score 0.12*LST region score. Alternatively, the LOH region score, TAI region score, and LST region score can be combined as follows to detect BRCA1/2 deficiency and/or predict response to therapy (e.g., response to platinum therapy, e.g., cisplatin): Combined AC region score = Arithmetic mean of LOH region score, TAI region score, and LST region score.

BRCA deficiency status and HRD status can be combined to predict response to therapy.

For example, the disclosure may include a method for predicting a patient's response (e.g., a patient with triple-negative breast cancer) to a cancer treatment regimen comprising a DNA-damaging agent (e.g., a platinum agent, e.g., cisplatin), an anthracycline, a topoisomerase I inhibitor, radiation, and/or a PARP inhibitor, the method comprising:

to determine, in a cancer cell from a patient sample, the number of AC indicator regions (e.g., LOH indicator regions, TAI indicator regions, LST indicator regions, or any combination thereof) on at least one pair of human chromosomes from a cancer cell of that cancer patient;

to determine whether a cancer cell from a patient sample is deficient in BRCA1 or BRCA2 (e.g., deleterious mutation, high promoter methylation); and

to diagnose a patient in whose sample (a) said number of indicator AC regions is higher than a reference number or (b) there is a BRCA1 or BRCA2 deficiency, or both (a) and (b), who is more likely to respond to said cancer treatment regimen.

In some embodiments of the method of the invention, the test value is derived by calculating the arithmetic mean of the numbers of the indicator LOH regions, the indicator TAI regions, and the indicator LST regions in said sample as follows:

Test value

(Number of indicator LOH regions) (Number of indicator TAI regions) (Number of indicator LST regions) = 3

and said one plus reference values were derived by calculating the arithmetic mean of the numbers of indicator LOH regions, indicator TAI regions and indicator LST regions in samples from said reference population as follows:

Test value

(Number of indicator LOH regions) (Number of indicator TAI regions) (Number of indicator LST regions) = 3

This document discloses an in vitro method for predicting patient response to a cancer treatment regimen comprising a DNA-damaging agent, anthracycline, a topoisomerase I inhibitor, or a PARp inhibitor, the method comprising:

(1) determine, in a sample comprising a cancer cell, the number of AC indicator regions comprising lOh indicator regions, TAI indicator regions and LST indicator regions in at least one pair of human chromosomes from a cancer cell of said patient;

(2) combine said AC indicator regions to provide a test value as follows: Test value = (number of LOH indicator regions) + (number of TAI indicator regions) + (number of LST indicator regions); and

(3) provide a reference value for comparison against said test value.

The reference value may represent the 5th percentile of the AC indicator region scores in a training cohort of HDR-deficient patients.

For illustrative purposes, this reference value is 42.

The method may also include comparing the test value with the reference value.

Additionally, the method may also include diagnosing a patient in whose sample the test value is greater than the reference value, having an increased probability of responding to the cancer treatment regimen.

In step (1) of the method, the determination of a combined number of indicator LOH regions, indicator TAI regions, and indicator LST regions in at least one pair of human chromosomes from the cancer cell may comprise assaying said sample to measure the copy number of each allele for at least 150, 200, 250, 300, 350, 400, 450, 500, 600, 700, 800, 900, 1,000, 1,500, 2,000, 2,500, 3,000, 3,500, 4,000, 4,500, 5,000, 6,000, 7,000, 8,000, 9,000, 10,000, 11,000, 12,000, 13,000, 14,000, 15,000, 16,000, 17,000, 18,000, 19,000, 20,000, 25,000, 30,000, 35,000, 40,000, 45,000, 50,000, 60,000, 70,000, 80,000, 90,000, 100,000, 125,000, 150,000, 175,000, 200,000, 250,000, 300,000, 400,000, 500,000, 600,000, 700,000 800,000, 900,000, 1,000,000 or more polymorphic genomic loci in at least 2, at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, at least 10, at least 11, at least 12, at least 13, at least 14, at least 15, at least 16, at least 17, at least 18, at least 19, at least 20, at least 21 or at least 22 autosomal pairs.

Optionally, this determination stage involves testing the sample to measure the number of copies of each allele for at least 5,000 polymorphic loci in these autosomal pairs.

Examples

Example 1 - LOH and TAI region scores in breast cancer subtypes and association with BRCA1/2 deficiency

A LOH signature based on genome-wide tumor LOH profiles has been developed that is highly correlated with defects in BRCA1/2 and other genes in the HDR pathway in ovarian cancer (Abkevich, et al., Patterns of Genomic Loss of Heterozygosity Predict Homologous Recombination Repair Defects, BR. J. CANCER (2012)), and that predicts response to therapy with DNA-damaging agents (e.g., neoadjuvant platinum-based) in breast cancer (Telli et al., Homologous Recombination Deficiency (HRD) score predicts response following neoadjuvant platinum-based therapy in triple-negative and BRCA1/2 mutation-associated breast cancer (BC), CANCER RES. (2012)). A second score based on the TAI score also shows a strong correlation with BRCA1/2 defects and predicts a response to platinum-based treatment in triple-negative breast cancer (Birkbaket et al., Telomeric allelic imbalance indicates defective DNA repair and sensitivity to DNA-damaging agents, CANCER Discov. (2012)). This study examined the frequency of BRCA1/2 defects and elevated LOH region or TAI scores in breast cancer subtypes defined by ER/PR/HER2 status.

Frozen tumors were acquired from three commercial tissue biobanks. For analysis, approximately 50 randomly selected tumors were chosen from each of four breast cancer subtypes (triple-negative, ER+/HER2-, ER-/HER2+, and ER+/HER2+). A custom hybridization panel targeting BRCA1, BRCA2, and 50,000 selected genome-wide SNPs was developed. This panel, along with sequencing on the Illumina HiSeq2500, was used to analyze the tumors for somatic and germline BRCA1/2 mutations, including large rearrangements and allelic SNP dosages. BRCA1 promoter methylation was determined by a cPCR assay (SA Biosciences). When available, normal tissue DNA was used to determine whether deleterious mutations were germline or somatic.

SNP data were analyzed using an algorithm that determines the most probable allele-specific copy number at each SNP location. The LOH region score was calculated by counting the number of LOH regions that were >15 Mb in length but shorter than the length of a complete chromosome. The TAI region score was calculated by counting the number of telomeric regions with allelic imbalance that were >11 Mb in length but did not span the centromere. Samples with poor-quality SNP data and/or high contamination with normal DNA were excluded. Of 213 samples, 191 yielded consistent scores.

Table 2: BRCA1/2 deficiency in IHC subtypes of breast cancer.

Table 3: Mutation screening was performed on compatible normal tissue from 17 of the BRCA1/2 mutants. 13 of the 17 individuals (765%) had a terminal line mutation.

T l 4: A i i n r l n i n L H TAI fi i n i BR A12

Figure 5 shows the scores of the LOH and TAI regions in the IHC subtypes of breast cancer.

5A: LOH score; 5B: TAI score. Blue bars: BRCA1/2 deficient samples. Red bars: unaltered BRCA1/2 deficient samples. Figure 6 shows the correlation between the LOH and TAI scores (correlation coefficient = 0.69). X-axis: LOH score; Y-axis: TAI score; red dots: unaltered samples; blue dots: BRCA1/2 deficient samples. The area under the dots is proportional to the number of samples with that combination of LOH and TAI scores (p = 10<-39>).

Logistic regression analysis was used to predict BRCA1/2 deficiency based on LOH and TAI scores. Both scores were significant in multivariate analysis (the x-value for LOH was 10.8 and for TAI 44.7; p = 0.001 and 2.3 × 10⁻¹¹). The best model for differentiating between deficient and unaffected BRCA1/2 samples was 0.32 × LOH region score ± 0.68 × TAI region score (p = 9 × 10⁻¹⁸).

Conclusions: Elevated scores in the LOH and TAI regions are strongly associated with BRCA1/2 deficiency in all breast cancer subtypes; the LOH and TAI region scores are significantly correlated; a combined AC region score (i.e., a combination of LOH and TAI) shows the strongest correlation with BRCA1/2 deficiency in this dataset. Based on this disclosure, the combination of LOH-HRD and TAI-HRD scores may predict response to DNA-damaging and other agents (e.g., platinum therapy) in triple-negative breast cancer, and may allow for the expansion of platinum use to other breast cancer subtypes.

Example 2- Scores of the LOH, TAI and LST regions in breast cancer subtypes and association with BRCA1/2 deficiency

SNP allele frequency ratios were obtained and used to calculate scores for the LOH, TAI, and LST regions as described in Example 1. The LST score was defined as the number of breakpoints between regions longer than 10 megabases that have a number of

to remove by filtering regions shorter than 3 megabases. It was observed that the LST score increased with ploidy in both unaltered and deficient samples. Therefore, instead of using ploidy-specific cutoff points in this Example 2, the LST region score was modified by adjusting for ploidy: LSTm = LST - kP, where P is ploidy and k is a constant. Based on a multivariate logistic regression analysis with deficiency as an outcome and LST and P as independent variables, k = 15.5.

Of 214 samples, 191 yielded scores that met the CC criteria used. Of these samples, 38 were BRCA1/2 deficient. The corresponding p-values according to the Kolmogorov-Smirnov test are: 8 × 10⁻¹² for the LOH region score, 2 × 10⁻¹⁶ for the TAI region score, and 8 × 10⁻⁸ for the LST region score. Of the 191 samples, 53 were triple-negative breast cancer, including 22 that were BRCA1/2 deficient. The corresponding p-values were 6 × 10⁻⁶, 3 × 10⁻⁶, and 0.0002 for the LOH, TAI, and LST region scores, respectively. When the same analysis is performed for each individual breast cancer subtype, significant p-values are also observed for all subtypes with at least one of the scores (Table 5). Figure 7A-C shows the distribution of scores for BRCA1/2 deficient samples versus BRCA1/2 unaffected samples.

The scores were then analyzed to determine if they were correlated (Figure 2D-F). The correlation coefficient between the LOH region score and the TAI region score was 0.69 (p = 10-39), between LOH and LST it was 0.55 (p = 2*10-19), and between TAI and LST it was 0.39 (p = 10-9).

To predict BRCA1/2 deficiency, a logistic regression analysis was used based on LOH, TAI, and LST scores. All three scores were significant in a multivariate analysis (the χ² value for LOH was 5.1 (p = 0.02), for TAI 44.7 (p = 2 × 10⁻¹¹), and for LST 5.4 (p = 0.02)). In this dataset, the best model for differentiating between deficient and unaffected BRCA1/2 samples was 0.21 × LOH, 0.67 × TAI, 0.12 × LST (p = 10⁻¹⁸). This Example 2 extends the findings of Example 1 (i.e., a model combining scores from the LOH and TAI regions) to a model combining scores from the LOH, TAI, and LST regions.

Other clinical data available for many of the samples included stage, grade, and age at diagnosis. Stage information was available for 64/191 samples. The correlation coefficient between stage and LOH region score (0.07) and TAI region score (0.1) was not significant. Grade information was available for 164/191 samples. The correlation coefficient between grade and LOH region score (0.33) and TAI region score (0.23) was significant (p = 2 × 10⁻⁵ and 0.004, respectively). Age at diagnosis was known for 184/191 samples. The correlation coefficient between age and LOH region score (-0.13) was not significant. The correlation coefficient between age and TAI region score (-0.25) was significant (p = 0.0009).

Table 5

continued

Example 3 - Arithmetic mean of the scores of the LOH, TAI, and LST regions in breast cancer subtypes and association with BRCA1/2 deficiency

The following study demonstrates how HRD scores, as described herein, can predict BRCA1/2 deficiency and the efficacy of HR-targeting agents in triple-negative breast cancer (TNBC). To investigate the rate of BRCA1/2 deficiency in breast cancer subtypes, breast tumor samples were analyzed for BRCA1/2 mutations and promoter methylation. The three HRD scores described in Example 2 were determined for these samples, and their association with BRCA1/2 deficiency was then examined using the arithmetic mean of the LOH/TAI/LST scores. A neoadjuvant cisplatin-treated TNBC cohort was further analyzed to examine the relationship between the three HRD scores and response.

Invasive breast tumor samples and compatible normal tissue samples were obtained from three commercial providers. Samples were selected to obtain approximately equal numbers of all breast cancer subtypes defined by immunohistochemical (IHC) analysis of ER, PR, and HER2. BRCA1 promoter methylation analysis was performed by cPCR. BRCA1/2 mutation screening and genome-wide SNP profiles were generated using a custom Agilent SureSelect XT capture followed by sequencing on an Illumina HiSeq2500. These data were used to calculate HRD-LOH, HRD-TAI, and HRD-LST scores.

SNP microarray and clinical data were downloaded from a public repository of cisplatin-1 and cisplatin-2 trial cohorts. BRCA1/2 mutation data from one of these cohorts were unavailable. The three HRD scores were calculated using publicly available data and analyzed for their association with cisplatin response. The two cohorts were combined to improve their statistical power.

To calculate HRD scores, SNP data were analyzed using an algorithm that determines the most probable allele-specific copy number at each SNP location. The HRD-LOH score was calculated by counting the number of LOH regions > 15 Mb in length but shorter than the length of a full chromosome. The HRD-TAI score was calculated by counting the number of regions > 11 Mb in length with allelic imbalance extending to one of the subtelomeres but not across the centromere. The HRD-LST score was the number of breakpoints between regions longer than 10 Mb after filtering out regions shorter than 3 Mb.

The combined score was the arithmetic mean of the LOH/TAI/LST scores. All p-values were derived from logistic regression models with BRCA deficiency or cisplatin response as the dependent variable.

Table 6 shows the BRCA1/2 mutation and BRCA1 promoter methylation frequency in four breast cancer subtypes. BRCA1/2 variant analysis was satisfactory in 100% of the samples, while large rearrangement analysis was less consistent, with 198/214 samples yielding data that exceeded the CC metrics. Deliriant mutations were observed in 24/214 individuals (one had a somatic mutation in BRCA1 and a germline mutation in BRCA2). Matched normal DNA was available for 23/24 mutants and was used to determine whether the identified mutation was germline or somatic. BRCA1 promoter methylation analysis was satisfactory in 100% of the samples. Figure 9 illustrates the HRD scores in BRCA1/2-deficient samples.

Table 6

Table 7 shows the association between the three HRD scores and BRCA1/2 deficiency in the entire breast cancer cohort. The combined score was the arithmetic mean of the three HRD scores.

Table 7

Table 8 shows the association between HRD scores and pCR (Miller-Payne 5) in TNBC treated with cisplatin in a neoadjuvant setting. Data were available from trial samples with cisplatin-1 (Silver et al., Efficacy of neoadjuvant cisplatin in triple-negative breast cancer. J. CLIN. ON10.28:1145-53 (2010)) and cisplatin-2 (Birkbak et al., (2012)). pCR was defined as Miller-Payne 5 status after neoadjuvant treatment. The pooled HRD score was the arithmetic mean of the three HRD scores.

Table 8

Conclusions: BRCA1/2 deficiency and elevated HRD scores were observed in all breast cancer subtypes, and the HRD score detected BRCA1/2 deficiency. All three HRD scores predicted/detected response to cisplatin treatment in TNBC. The average of the three HRD scores (arithmetic mean) detected BRCA1/2 status in a full breast cancer cohort and response to cisplatin in a second, independent TNBC cohort. The combined HRD mean was a stronger independent predictor of BRCA1/2 deficiency or response to therapy than the individual HRD scores.

Example 4 - Multivariate analysis of BRCA1/2 status and DNA-based assays regarding homologous recombination deficiency

The previous examples described DNA-based scores that measure homologous recombination deficiency (HRD), demonstrating that each score is significantly associated with BRCA1/2 deficiency, as is a pooled HRD score defined as the arithmetic mean of three different HRD scores. This example extends the results of the previous examples by examining (1) associations between each of the three scores and the pooled HRD score, (2) associations of clinical variables with the pooled HRD score, and (3) associations of clinical variables and the pooled HRD score with BRCA1/2 deficiency.

Methods: In this Example 4, the analyses include the same 197 patient samples described in the previous Examples. In summary, 215 breast tumor samples were acquired as fresh frozen specimens from 3 commercial suppliers. The samples were selected to provide approximately equal representation of breast cancer subtypes based on IHC analysis of ER, PR, and HER2. According to a Kolmogorov-Smirnov quality metric, 198 samples yielded reliable HRD scores. One patient with a transient HRD score was excluded from the analysis due to an unusual breast cancer subtype (ER/PR+ HER2-). The patient's tumor and clinical characteristics are detailed in Table 9.

Clinical data were provided for 91 patient variables, but data for most variables were too scarce for inclusion in the analysis. Breast cancer subtype (TNBC, ER+/HER2-, ER-/HER2+, ER+/HER2+) was available for all patients. Other variables considered were age at diagnosis (provided for 196/197 patients), stage (provided for 191/197 patients), and grade (provided for 190/197 patients).

BRCA1/2 mutation screening and genome-wide SNP profiling were generated using a custom Agilent SureSelect XT capture followed by sequencing on an Illumina HiSeq2500. BRCA-1 promoter region methylation was determined by cPCR. Samples with methylation greater than 10% were classified as methylated.

HRD scores were calculated from loss of heterozygosity (LOH) profiles (HRD-LOH), telomeric allelic imbalance (HRD-TAI), and large-scale state transitions (HRD-LST) across the entire tumor genome. The three HRD scores combined into the "combined HRD score" are discussed in this Example 4.

BRCA1/2 deficiency was defined as loss of function resulting from a BRCA-1 or BRCA-2 mutation, or methylation of the BRCA-1 promoter region, along with loss of heterozygosity (LOH) in the affected gene.

All statistical analyses were performed using R version 3.0.2. All reported p-values were two-tailed. Statistical tools employed included Spearman's rank correlation coefficient, one-way Kruskal-Wallis analysis of variance, and logistic regression.

For the logistic regression model, HRD scores and age at diagnosis were coded as numerical variables. Breast cancer stage and subtype were coded as categorical variables. Grade was analyzed as both a numerical and categorical variable, but was categorical unless otherwise noted. Coding grade as a numerical variable is not appropriate unless the increase in BRCA1/2 deficiency likelihood ratios is the same when comparing grade 2 patients with grade 1 patients, as well as when comparing grade 3 patients with grade 2 patients.

The p-values reported for univariate logistic regression models are based on the partial likelihood ratio. Multivariate p-values are based on the partial likelihood ratio with respect to the change in deviation of a full model (which includes all relevant independent variables) versus a reduced model (which includes all independent variables except the independent variable being evaluated, and any interaction terms that affect the independent variable being evaluated). Odds ratios for HRD scores are reported as interquartile ranges.

Results: Pairwise correlations of HRD-LOH, HRD-TAI, and HRD-LST scores were examined graphically (Figure 1) and quantified using Spearman's rank-sum correlation. Spearman's rank-sum correlation was preferred over Pearson's product-moment correlation, which is more commonly used, because skewness and outliers were observed in the HRD score distributions. All pairwise score comparisons showed a significantly positive correlation different from zero (p < 10.16).

The extent of independent information on BRCA1/2 deficiency captured by each of the HRD-LOH, HRD-TAI, and HRD-LST scores was measured by examining a multivariate logistic regression model with all three scores included as independent variables for BRCA1/2 deficiency status (Table 10). The HRD-TAI score captured significant information on BRCA1/2 deficiency independent of that provided by the other two scores (p=0.00016), as did the HRD-LST score (p=0.00014). At the 5% significance level, the HRD-LOH score did not add significant independent information on BRCA1/2 deficiency (p=0.069).

Table 10

Table 10 illustrates the results of a 3-term multivariate logistic regression model with HRD-LOH, HRD-TAI, and HRD-LST as independent variables for BRCA1/2 deficiency.

To assess whether the combined HRD score adequately captured BRCA1/2 deficiency information from its three components, three bivariate logistic regression models were tested. Each model included the combined HRD score and one of the HRD-LOH, HRD-TAI, or HRD-LST scores. None of the component scores contributed significantly to the combined HRD score at the 5% significance level (HRD-LOH p = 0.89, HRD-TAI p = 0.090, HRD-LST p = 0.28). This suggests that the combined HRD score adequately captures BRCA1/2 deficiency information from the HRD-LOH, HRD-TAI, and HRD-LST scores.

The combined HRD score was then compared to a model-based composite score optimized to predict BRCA1/2 deficiency in this patient group. While the combined HRD score weights each of the HRD-LOH, HRD-TAI, and HRD-LST scores equally, the model-based score assigns approximately twice the weight to the HRD-TAI score as to the HRD-LOH or HRD-LST scores. The formula for the model-based score is as follows:

HRD based on the model = 0.11 X (HRD-LOH) 0.25X (HRD-TAI) 0.12 X (HRD-LST).

The results of the univariate analysis (Table 11) show that the model-based HRD score surpasses the combined HRD score by approximately one order of magnitude (model-based HRD p=2.5x10'25, combined HRD p=1.1x10'24).

Table 11

Table 11 shows the results of univariate logistic regression. The likelihood ratios for HRD scores are recorded as the IQR of the score. The likelihood ratio for age is recorded in years. The likelihood ratio for grade (numerical) is per unit.

In a bivariate logistic regression model, the model-based HRD score did not add significant independent information about BRCA1/2 deficiency compared to the combined HRD score (p = 0.089). This further suggests that the combined HRD score adequately captures information about BRCA1/2 deficiency from the HRD-LOH, HRD-TAI, and HRD-LST scores.

Figure 12 shows associations of clinical variables with the combined HRD score. The combined HRD score was significantly correlated with tumor grade (Spearman correlation 0.23, p = 0.0017). Correlations with breast cancer stage and age at diagnosis were not significantly different from zero at the 5% level. Mean combined HRD scores differed significantly among breast cancer subtypes (p = 1.6 x 10⁵) according to a one-way Kruskal-Wallis analysis of variance test.

The heterogeneity of the combined HRD score among clinical subpopulations was tested by examining the significance of interaction terms in multivariate logistic regression models. For each clinical variable, an interaction term with the combined HRD score was added to a model that included all clinical variables and the combined HRD score. None of the interaction terms reached significance at the 5% significance level. Therefore, there is no evidence to suggest that the probability of BRCA1/2 deficiency conferred by the combined HRD score varies among clinical subpopulations.

Analogous tests for each of the HRD-LOH, HRD-TAI, and HRD-LST scores indicated a significant interaction of the HRD-TAI score with age (p = 0.0072) and grade (p = 0.015), and a significant interaction of the HRD-LST score with breast cancer subtype (p = 0.021). Adjusted for multiple comparisons, only the interaction of the HRD-TAI score with age maintained significance at the 5% level (p = 0.029). The significance of this interaction suggests that the increased likelihood of BRCA1/2 deficiency per unit increase in the HRD-TAI score decreases with increasing age.

Figure 13 shows associations of clinical variables with BRCA1/2 deficiency. Clinical variables and the combined HRD score were assessed using univariate (Table 11) and multivariate (Table 12) logistic regression models. HRD score likelihood ratios are reported as IQRs. Likelihood ratios for age at diagnosis are reported in years.

Table 12

Table 12 shows the results of multivariate logistic regression. The likelihood ratios for HRD scores are reported as IQR of the score. The likelihood ratio for age is reported in years.

In univariate analysis, each of the HRD scores (HRD-LOH, HRD-TAI, HRD-LST, combined HRD, and model-based HRD) was significantly associated with BRCA1/2 deficiency. Higher scores indicated a greater likelihood of deficiency. Increasing age at diagnosis was significantly associated with a lower risk of BRCA1/2 deficiency (p = 0.0071). Univariate analysis results for breast cancer subtype and tumor grade (both categorical and numerical) were also statistically significant. Cancer stage was not associated with BRCA1/2 status.

In multivariate analyses, a model based on the combined HRD score and all available clinical variables was examined. The combined HRD score captured significant information about BRCA1/2 deficiency that was not captured by the clinical variables (p = 1.2 x 10⁻¹⁶). Of the available clinical variables, only age at diagnosis remained significant in the multivariate setting (p = 0.027). Grade, coded as a categorical variable, was not statistically significant (p = 0.40). Grade was also not significant when coded as a numerical variable (p = 0.28). The quadratic and cubic effects of the combined HRD score were tested in multivariate models that included all clinical variables, but these were not statistically significant.

Argument. In this Example 4, the frequency of BRCA1/2 defects ranged from ~9% to ~16% in four breast cancer subtypes, as defined by IHC subtyping. Sequencing of normal and compatible tumor DNA samples suggests that approximately 75% of the observed mutations were germline in origin. The primary method for detecting loss of the second allele in breast cancer is through LOH; however, approximately 24% of the tumors also carried subsequent deleterious somatic mutations in the second allele. In addition, an apparently sporadic breast tumor was observed in an individual carrying a deleterious somatic BRCA2 mutation.

The three HRD scores showed a strong correlation with BRCA1/2 deficiency regardless of subtype, and the frequency of elevated scores suggests that a significant proportion of all breast tumor subtypes have defects in the DNA repair pathway by homologous recombination. These findings, especially when combined with those in Example 3 above, show that agents that target DNA or exploit DNA damage repair (e.g., platinum agents) may prove effective in a subset of tumors (those with homologous recombination deficiency detected in accordance with this disclosure) across all breast cancer subtypes.

When applying these HRD scores, individually or in combination, in a clinical setting, it is best to use an assay compatible with formalin-fixed and paraffin-embedded (FFPE) core needle biopsies. Samples of this type yield very little and low-quality DNA. DNA extracted from these FFPE-treated samples typically does not perform well in SNP microarray analysis.

Target enrichment technologies based on liquid hybridization have been developed for the production of libraries intended for next-generation sequencing. These methodologies allow targeted sequencing of regions of interest after genomic complexity reduction, resulting in lower sequencing costs. Preliminary tests indicated that available assays were compatible with DNA derived from FFPE DNA. In Example 4, we report the development of a capture panel targeting approximately 54,000 SNPs distributed throughout the genome. The allele counts from the sequencing information provided by this panel can be used for copy number and λOμ reconstruction, and for calculating the three HRD scores. Furthermore, BRCA1 and BRCA2 capture probes can be included in the panel, as in Example 4, enabling high-quality mutation screening for deleterious variants in these genes in the same assay.

The three scores were significantly correlated with each other, suggesting that they all measure the same core genomic phenomenon. However, logistic regression analysis indicates that the scores could be combined, resulting in a stronger association with BRCA1/2 deficiency in this dataset.

The combination of a consistent scoring system capable of identifying tumors with defects in homologous recombination DNA repair and an assay compatible with formalin-fixed, paraffin-embedded clinical pathology specimens facilitates the diagnostic identification and classification of patients with a high probability of response to agents that target double-stranded DNA repair. Furthermore, such agents may be useful in all breast cancer subtypes in which HRD is detected according to this disclosure.

Example 5: High threshold value of HRD (e.g., an example of an HRD signature)

This example demonstrates the determination of a high HRD. A highly sensitive reference threshold value was selected to detect HRD in breast and ovarian tumors that was not specific to treatment response or outcome. The total number of LOH, TAI, and LST regions was determined. To calculate HRD scores, SNP data were analyzed using an algorithm that determines the most probable allele-specific copy number at each SNP location. The HRD-LOH score was calculated by counting the number of LOH regions with a length > 15 Mb, but shorter than the length of a complete chromosome. The HRD-TAI score was calculated by counting the number of regions with a length > 11 Mb with allelic imbalance extending to one of the subtelomeres, but not across the centromere. The HRD-LST score was the number of breakpoints between regions longer than 10 Mb after removing regions shorter than 3 Mb by filtering. The combined score (HRD score) was the sum of the LOH/TAI/LST scores.

The training set was assembled from four different cohorts (497 breast cancer cases and 561 ovarian cancer cases). The set consisted of 78 breast tumors and 190 ovarian tumors that lacked a functional copy of BRCA1 or BRCA2, because the distribution of HRD scores in BRCA-deficient samples reflects the distribution of scores in HRD samples overall. The threshold was set at the 5th percentile of HRD scores in the training set, providing >95% sensitivity for detecting HR deficiency. A high HRD (or HRD signature) was defined as having a reference score >42 (Figure 14).

Example 6 - HRD predicts response to cisplatin in triple-negative breast cancer

This example demonstrates how HRD scores, as described herein, can predict the efficacy of agents targeting HR deficiency in triple-negative breast cancer (TNBC) samples. We examined the relationship between the three HRD scores and response in a neoadjuvant TNBC cohort treated with cisplatin. All values were derived from logistic regression models with cisplatin response as the dependent variable.

The HR deficiency status was determined in 62 of the 70 samples (70 individual patients) received from a cisplatin-treated cohort (8 had insufficient tumors for analysis). Of these, 31 (50%) were HR-deficient, 22 (35%) were not HR-deficient, and 9 (15%) were indeterminate. Figure 15 provides a histogram showing the distribution of HRD scores in the cohort. Scores > 42 were considered to indicate high HRD (see also Example 5). The bimodality illustrated in Figure 15 indicates that HRD scores effectively distinguish between HR-deficient and non-deficient tumor states. Pathological complete response (pCR), which is associated with prolonged survival, was defined as a residual cancer burden (RCB) of 0 and was observed in 11/59 (19%) samples. Pathological response (PR) was defined as a CRC of 0 or 1 and was observed in 22/59 (37%) samples. These overall response rates correlated with monotherapy expectations.

The statistical analyses followed a predefined Statistical Analysis Plan (SAP), which included primary, secondary, and natural-type BRCA subset analyses.

The primary analysis used HR deficiency status to predict response in 50 samples. As shown in Table 13, HR-deficient samples provided a better independent variable for response for both PR and PRc. For example, 52% of HR-deficient samples had a pathological response compared to 9.5% of non-deficient samples that had a pathological response. Similarly, 28% of HR-deficient samples had a complete pathological response compared to 0% of non-deficient samples that had a complete pathological response.

T l 1 : An li i rim r iliz n fi i n i HR r r ir n r

The secondary analysis used a quantitative HRD score, as described in Example 5, to predict response in 48 samples. As shown in Table 14, HRD scores were significantly higher in responding patient samples compared to non-responders, defined as PR or CRp.

T l 14: An li i n ri iliz n n i n HRD n i iv r r ir l r

The distribution of HRD scores in each response class of the secondary analysis, as defined by BRCA mutation status, is illustrated in Figure 16, where the dotted line at value 42 represents the HRD threshold between low and high scores. The response curve, or the probability of PR associated with each quantitative HRD score value from the secondary analysis, is illustrated in Figure 17. The curve shown in Figure 17 was modeled using generalized logistic regression, which determines four parameters: shape, scale, and the lower and upper bounds of the curve. Shaded cells indicate the probability of response in samples with homologous recombination (HR) deficiency versus samples without such deficiency. Table 15 shows that in the secondary analysis, HR status remained significantly associated with pathological response.

T l 1 : M l m l iv ri n r l i

continued

Table 16 shows scores for the individual HRD components versus pathological response and is illustrated in Figure 18. Table 16 shows that each component score—LOH, TAI, and LST—was predictive of response, and their sum, the HRD score, was equal to or more significant than any of the individual components (HRD p-value = 3.1 x 10⁻⁴). Figure 18 illustrates strong pairwise correlations between the component scores.

Table 16: Quantitative scores of the HRD component versus RP

In the secondary analysis, the association of BRCA1/2 mutation status with response was also verified. Table 17 confirmed that BRCA mutation status was associated with response; however, the association was not significant in this cohort (n=51), and BRCA mutation status was not as predictive as HR deficiency.

Furthermore, a subset analysis was performed using RH Deficiency status in 38 natural-type BRCA samples to demonstrate that RH Deficiency is predictive in BRCA1/2 mutation-free samples. As shown in Table 18, HR-deficient samples provided a better independent variable for response to both PR and cPR in natural-type BRCA samples. For example, 52.6% of HR-deficient samples had a pathological response compared to 10.5% of non-deficient samples that had a pathological response. Similarly, 26.3% of HR-deficient samples had a complete pathological response compared to 0% of non-deficient samples that had a complete pathological response.

Table 18: Subset analysis using HR deficiency to predict response in BRCA samples of ti or natural

A subset analysis was also performed using the quantitative HRD score in 38 natural-type BRCA samples. As shown in Table 19, samples with high HRD (scores > 42) provided a better independent response variable for both PR and pCR in natural-type BRCA samples.

Table 19: Subset analysis using quantitative HRD scores to predict response in natural-type BRCA samples

In conclusion, this example demonstrates that the sum of the three HRD scores significantly predicted the response to cisplatin treatment in TNBC.

Claims (16)

1. An in vitro method for determining the homologous recombination deficiency (HR) status of a patient cancer cell, comprising: (1) determine, in a sample comprising the patient's cancer cells, a combined number of Indicator LOH regions, indicator TAI regions, and indicator LST regions in at least one pair of human chromosomes of the cancer cell; (2) calculate a test value derived from the combined number of indicator LOH regions, regions of TAI indicators and LST indicator regions, where the test value is derived using a formula selected from: (Yo) Test value = (Number of indicator LOH regions) (Number of indicator TAI regions) (Number of indicator LST regions) 3 (ii) Test value = (LOH region score) (TAI region score) + (LST region score), (iii) Test value A * (n.B of LOH indicator regions) B * (n.B of TAI indicator regions) C * (n.B of LST indicator regions) =3 and (iv) Test value = A * (LOH region score) B * (TAI region score) C *(LST region score); and (3) Identify the cancer cell with HR deficient status when the test value exceeds a reference value, where the reference value is 32 or higher. 2. The method of claim 1, wherein the test value is an arithmetic mean of the number of regions of LOH indicators, TAI indicator regions and LST indicator regions. 3. The method of claim 1, wherein the reference value is 40 or greater, or optionally 42 or greater. 4. The method of any one of claim 1 or 3, wherein the test value is derived using a selected formula from: Test value = (LOH region score) (TAI region score) + (LST region score), and Test value = A * (LOH region score) B * (TAI region score) C *(LST region score). 5. The method of any one of claims 1 to 3, wherein the test value is derived using a selected formula from: Test value (Number of indicator LOH regions) (Number of indicator TAI regions) (Number of indicator LST regions) = 3 and Test value A * (n.s of LOH indicator regions) B * (n.B of TAI indicator regions) + C * (n.B of LST indicator regions] = 3 6. The method of any one of claims 1-5, wherein the indicator LOH regions are more longer than 1.5 megabases but shorter than the entire length of the respective chromosome within which they are located find the LOH region. 7. The method of claim 6, wherein the indicator LOH regions are at least 14 megabases of length, or optionally where the indicator LOH regions are at least 15 megabases in length. 8. The method of any one of claims 1-7, wherein the indicator TAI regions are regions with allelic imbalance that (i) extend to one of the subtelomeres, (ii) do not cross the centromere, and (iii) are longer than 1.5 megabases, optionally where the indicator TAI regions are at least 10 megabases in length. 9. The method of any one of claims 1-8, wherein the indicator LST regions are regions comprising a somatic copy number breakpoint along a chromosome that is located between two regions of at least 10 megabases in length after removing shorter regions by filtering that 3 megabases in length. 10. The method of any one of claims 1-9, wherein the at least one pair of human chromosomes They are autosomes. 11. The method of any one of claims 1-9, wherein the human chromosomes are autosomes and where the combined number of LOH indicator regions, TAI indicator regions, and LST regions Indicators are determined in at least 10 or optionally at least 15 pairs of autosomes. 12. The method of any one of claims 10-11, further comprising testing at least 150 polymorphic genomic loci in each autosomal pair. 13. The method of any one of claims 1-9, further comprising testing at least 5,000 polymorphic genomic loci on at least 20 pairs of human chromosomes, where the chromosomes are autosomes. 14. The method of any one of claims 1-13, further comprising identifying the patient as likely to respond to a cancer treatment regimen comprising a DNA-damaging agent, anthracycline, topoisomerase I inhibitor, or PARP inhibitor based on cell identification cancerous as being probably HR deficient, optionally where the DNA-damaging agent is cisplatin, carboplatin, oxaliplatin or picoplatin, the anthracycline is epirubincin or doxorubicin, the topoisomerase I inhibitor is campothecin, topotecan or irinotecan, or the PARP inhibitor is iniparib, olaparib or velaparib. 15. The method of any one of claims 1-13, further comprising identifying the patient as likely to respond to a cancer treatment regimen comprising a DNA-damaging agent, anthracycline, topoisomerase I inhibitor, or PARP inhibitor based on cell identification cancerous as being probably HR deficient, optionally which also includes identifying the patient as likely to respond to a regimen of treatment comprising one or more taxane agents, a growth factor, or a receptor inhibitor of the growth factor, or an antimetabolite agent based on the identification of the cancer cell as probably not HR deficient, optionally where the taxane agent is doxetaxel, paclitaxel, abraxane, the growth factor or The growth factor receptor inhibitor is erlotinib, gefitinib, lapatinib, sunitinib, bevacizumab, cetuximab, trastuzumab, panitumumab, or where the antimetabolite is 5-fluorouracil or methotrexate. 16. The method of any one of claims 1-15, wherein the cancer cell is a cancer cell ovarian cancer, a breast cancer cell, a liver cancer cell, an esophageal cancer cell, a lung cancer cell, a head and neck cancer cell, a prostate cancer cell, a cell of colon, rectal or colorectal cancer or a pancreatic cancer cell.